3. Basic Operations

YottaDB Environment Setup

Several environment variables control the operation of YottaDB. Some of them are required for normal operation, whereas for others, YottaDB assumes a default value if they are not set. The three most important ones are:

  • ydb_dist: Points to the directory where YottaDB is installed. If it is not set, the yottadb program sets it to the directory from which it is executed. Processes started with the JOB command inherit this and other environment variables from their parent processes.

  • ydb_gbldir: Points to a global directory file. A global directory file maps global variables (the hierarchical key-value relationships of the YottaDB database) to database files in the filesystem. This is not required by processes that do not access any global variables. There is no default.

  • ydb_routines: Tells a process where to find application code written in M. It is not required by applications that do not invoke M code. If it is not set, the yottadb program sets it to $ydb_dist/libyottadbutil.so where YottaDB's own utility routines (such as GDE) are available.

A YottaDB distribution comes with many scripts that set up a default YottaDB environment for the shell of your choice. These scripts are as follows:

  • ydb_env_set (gtmprofile): uses reasonable defaults to set up a YottaDB application development environment that includes plugins conforming to the YottaDB standard (such as the POSIX plugin and Octo). The ydb_env_set file when sourced sets default values for the environment variables ydb_dir, ydb_dist, ydb_etrap, ydb_gbldir, ydb_log, ydb_readline, ydb_rel, ydb_repl_instance, ydb_retention, ydb_routines, ydb_tmp, gtmdir, gtm_dist, gtm_etrap, gtmgbldir, gtm_log, gtm_repl_instance, gtm_retention, gtmroutines, gtm_tmp, and gtmver. Additionally, if $ydb_chset is not set, or set to a case independent "UTF-8", sourcing the file also sets ydb_icu_version and gtm_icu_version. The default environment is created under $ydb_dir, defaulting to $HOME/.yottadb if ydb_dir is undefined. It also deletes prior generation journal and temporary files older than the number of days specified by the environment variable ydb_retention. It attempts to automatically recover the database when it runs and as such is suitable for "out of the box" usage of YottaDB. Although it will work for large multi-user environments, you may want to modify or replace it with custom scripting tailored to your specific needs. Sourcing ydb_env_set saves environment variables that it sets such that subsequently sourcing ydb_env_unset restores the saved values.

  • ydb_env_unset: unsets environment variables set by a prior ydb_env_set in the same shell session, and restores prior values, if any.

  • ydb (gtm): The ydb script starts YottaDB after sourcing ydb_env_set.

  • gdedefaults: a GDE command file that specifies the default values for database characteristics defined by GDE.

These scripts are designed to give you a friendly out-of-the-box YottaDB experience. Even though you can set up an environment for normal YottaDB operation without using these scripts, going through these scripts will help you understand how to configure environments.

ydb_env_set

On POSIX shells, ydb_env_set manages a basic YottaDB environment, and sets reasonable values for environment variables for YottaDB operation:

ydb_dir, ydb_dist, ydb_etrap, ydb_gbldir, ydb_log, ydb_readline, ydb_rel, ydb_repl_instance, ydb_retention, ydb_routines, ydb_tmp, gtmdir, gtm_dist, gtm_etrap, gtmgbldir, gtm_log, gtm_repl_instance, gtm_retention, gtmroutines, gtm_tmp, gtmver

You can set the following environment variables before sourcing ydb_env_set or running the ydb script;

  • ydb_chset - set it to "M" to run YottaDB in M mode and "UTF-8" to run YottaDB in UTF-8 mode. The default when sourcing ydb_env_set is UTF-8 mode. YottaDB natively (i.e., without sourcing ydb_env_set) defaults to M mode. Since UTF-8 mode requires the locale setting LC_CTYPE (as reported by the locale command) to specify a character set with UTF-8 encoding, if a locale is not specified, ydb_env_set also attempts to set an available UTF-8 locale. As YottaDB in UTF-8 mode may require ydb_icu_version to be set, if it is not set, ydb_env_set attempts to set the environment variable to the version installed on the system.

  • ydb_dir - set this to define a directory for the environment set and managed by ydb_env_set. If it is not set, ydb_env_set uses $HOME/.yottadb.

To use Database Encryption, set the ydb_passwd and ydb_crypt_config environment variables first and then source ydb_env_set.

The following shell variables are used by the script and may have values upon its completion:

old_ydb_dist, old_ydb_routines, old_gtmver, tmp_ydb_tmp, tmp_passwd.

The $ydb_routines value set by the ydb_env_set script enables auto-relink by default for object files in the $ydb_dir/$ydb_rel/o directory in M mode and $ydb_dir/$ydb_rel/o/utf8 in UTF-8 mode, where $ydb_rel is the current release and platform, e.g., r1.36_x86_64. Auto-relink requires shared memory resources and limits beyond those for database operation. If your system has inadequate shared memory configured, YottaDB displays messages along the lines of:

%YDB-E-SYSCALL, Error received from system call shmget() failed

Refer to your OS documentation to configure shared memory limits (for example, on common Linux systems, the kernel.shmmax parameter in /etc/sysctl.conf).

ydb_env_set ensures that ydb_dist is set correctly.

When ydb_env_set is sourced, it provides a default execution environment (global directory and a default multi-region database) if none exists, and recovers an existing environment if needed. If neither $ydb_dir nor $gtmdir are defined, by default it creates the database in $HOME/.yottadb. The structure of the default environment is shown below, with routines in the r subdirectories, object code in the o and o/utf8 subdirectories. Databases are in the g subdirectory.

.yottadb
├── r
├── r1.36_x86_64
│   ├── g
│   │   ├── %ydbaim.dat
│   │   ├── %ydbaim.mjl
│   │   ├── %ydbocto.dat
│   │   ├── %ydbocto.mjl
│   │   ├── yottadb.dat
│   │   ├── yottadb.gld
│   │   └── yottadb.mjl
│   ├── o
│   │   └── utf8
│   └── r
└── V6.3-014_x86_64 -> r1.36_x86_64

In M mode, ydb_env_set sets the environment variable ydb_routines to something like the following (where $ydb_dist and $ydb_rel are as discussed above):

$ydb_dir/$ydb_rel/o*($ydb_dir/$ydb_rel/r $ydb_dir/r) $ydb_dist/libyottadbutil.so

$ydb_dir/$ydb_rel/o*($ydb_dir/$ydb_rel/r $ydb_dir/r) specifies that for object code in $ydb_dir/$ydb_rel/o YottaDB searches for routines in $ydb_dir/$ydb_rel/r, then $ydb_dir/r. For routines not found there, it looks in $ydb_dist/libyottadbutil.so. The * suffix after the object directory enables the auto-relink facility.

For a comprehensive discussion of YottaDB source and object file management, refer to the $ZROUTINES section in the Programmer's Guide.

When $ydb_chset is set to UTF-8, ydb_env_set sets ydb_routines to something like this:

$ydb_dir/$ydb_rel/o/utf8*($ydb_dir/$ydb_rel/r $ydb_dir/r) $ydb_dist/utf8/libyottadbutil.so

If you have installed any plugins that include shared libraries, the ydb_env_set script includes those. For example, with the POSIX plugin installed, ydb_dir set to /home/jdoe1 and YottaDB installed in /usr/local/lib/yottadb/r136, ydb_env_set would set ydb_routines in M mode to:

/home/jdoe1/.yottadb/r1.36/o*(/home/jdoe1/.yottadb/r1.36/r /home/jdoe1/.yottadb/r) /usr/local/lib/yottadb/r136/plugin/o/_ydbposix.so /usr/local/lib/yottadb/r136/libyottadbutil.so

Note

While sourcing ydb_env_set provides reasonable defaults, please see environment variables for more finer-grained control of YottaDB configuration and operation.

ydb_env_set creates the following alias: alias gde="$ydb_dist/yottadb -run GDE"

Sourcing ydb_env_set manages a four region database:

  1. A DEFAULT region in which empty string ("") subscripts are disabled. All global variables except those mapped to the YDBAIM, YDBJNLF, and YDBOCTO regions, are mapped to the DEFAULT region.

  2. A YDBAIM region, intended to be used by the Application Independent Metadata plugin with the following properties:

    • Empty string subscripts are enabled.

    • ^%ydbAIM* global variables (with all combinations of capitalization of "AIM") are mapped to YDBAIM.

    • The key size is 1019 bytes and the record size is 1MiB.

    • The default database filename is $ydb_dir/$ydb_rel/g/%ydbaim.dat and the default journal file is $ydb_dir/$ydb_rel/g/%ydbaim.mjl.

    • The block size for YDBAIM is 2KiB, with an initial allocation of 10000 blocks, extended by 20000 blocks.

    • 2000 global buffers.

    Except for these differences, the properties of the YDBAIM region are the same as those of the DEFAULT region.

  3. A YDBJNLF region, intended to be used the %YDBJNLF utility routine with the following properties:

    • ^%ydbJNLF* global variables (with all combinations of capitalization of "JNLF") are mapped to YDBJNLF.

    • The key size is 1019 bytes and the record size is 1MiB.

    • The default database filename is $ydb_dir/$ydb_rel/g/%ydbjnlf.dat which has AUTODB set so that the database file is only created when needed or referenced. The %YDBJNLF utility program is most likely to be used for troubleshooting and forensics on systems other than those used for normal development, testing and production.

    • Journaling is not enabled by default, and not recommended, as discussed in the %YDBJNLF documentation.

    • The YDBJNLF region uses the MM access method.

    Except for these differences, the properties of the YDBJNLF region are the same as those of the DEFAULT region.

  4. A YDBOCTO region, intended to be used by the Octo SQL plugin with the following properties:

    • Empty string subscripts are enabled.

    • ^%ydbOcto* global variables (with all combinations of capitalization of "Octo") are mapped to YDBOCTO.

    • The key size is 1019 bytes and the record size is 1MiB.

    • The default database filename is $ydb_dir/$ydb_rel/g/%ydbocto.dat and the default journal file is $ydb_dir/$ydb_rel/g/%ydbocto.mjl.

    • The block size is 2KiB, with an initial allocation of 10000 blocks, extended by 20000 blocks.

    • 2000 global buffers.

    Except for these differences, the properties of the YDBOCTO region are the same as those of the DEFAULT region.

Additionally:

  • The default mode is UTF-8 if YottaDB was installed with UTF-8 support.

  • For UTF-8 mode, sourcing ydb_env_set checks whether a locale is specified in the LC_ALL or LC_CTYPE environment variables. If not, it uses the first UTF-8 locale in the locale -a output, and terminates with an error if one is not found.

  • In case of error, the location of the error file is output.

Sourcing ydb_env_set handles the case where replication is turned on.

ydb_env_set was modified in YottaDB effective release r1.36, to create a four region database.

Sourcing ydb_env_set respects existing values of ydb_gbldir and ydb_routines. Other features include:

  • ydb_routines creates environment variables to call exported C routines in the $ydb_dist/plugin directory.

  • In the event the instance is coming up after an unclean shutdown, it recovers the instance using MUPIP JOURNAL RECOVER or MUPIP JOURNAL ROLLBACK depending on the replication setting.

  • Global directory files are automatically upgraded to the format of the current YottaDB release.

Sourcing ydb_env_set sets $?:

  • 0 is normal exit

  • 1 means that $ydb_dist does not match the location of ydb_env_set

  • 2 means that $gtm_dist does not match the location of ydb_env_set

  • 3 means that neither $ydb_dist nor $gtm_dist match the location of ydb_env_set

  • Other non-zero codes are as returned by set^%YDBENV

ydb_env_set was modified to set $status in YottaDB effective release r1.34.

gdedefaults

Specifies default or template values for database characteristics for the environment managed by ydb_env_set. You should not normally need to edit it.

ydb

By default, ydbinstall creates a symbolic link to $ydb_dist/ydb from /usr/local/bin/ydb.

The ydb script starts with #!/bin/sh so it can be invoked from any shell. You can use it to both run a program and run YottaDB in direct mode. It sources ydb_env_set as described above. This script sets up everything you need to run YottaDB for a simple out-of-the-box experience, and should work for many multi-user environments. For larger systems, you can create an alternative script specific to your configuration.

To run the ydb script, type:

$ /usr/local/bin/ydb

To invoke help, type:

$ /usr/local/bin/ydb -help
ydb -dir[ect] to enter direct mode (halt returns to shell)
ydb -help / ydb -h / ydb -? to display this text
ydb -run <entryref> to start executing at an entryref
ydb -version to display version information

ydb -version to display version information

Environment Variables

YottaDB supports both ydb_* environment variables and GT.M environment variables (referred to here as gtm*, though some are upper case). If the ydb* environment variable is not defined, but the gtm* environment variable is is, the latter is used. If the ydb* environment variable and the gtm* environment variable are both defined, the ydb* environment variable value takes precedence.

Environment variables of the form ydb_xc_<package> are used to point to external call tables; the GT.M names of these variables are of the form GTMXC_<package>.

A comprehensive list follows of environment variables that are directly or indirectly used by YottaDB:

EDITOR

EDITOR is a standard system environment variable that specifies the full path to the editor to be invoked by YottaDB in response to the ZEDIT command (defaults to vi, if $EDITOR is not set).

LC_CTYPE

LC_CTYPE is a standard locale setting reported by the locale command, and which can be explicitly set through the LC_ALL or LC_CTYPE environment variables. When $ydb_chset has the value "UTF-8", LC_CTYPE reported by the locale command must specify a UTF-8 locale (e.g., "en_US.utf8").

LC_ALL

LC_ALL is a standard locale setting reported by the locale command, and which can be explicitly set in the environment. Setting LC_ALL also sets LC_TYPE and has a more pervasive effec than just setting LC_CTYPE.

LD_LIBRARY_PATH

LD_LIBRARY_PATH is a standard system environment variable used to modify the default library search path. Use this extension when YottaDB relies on custom compiled libraries that do not reside in the default library search path.

TZ

TZ is a standard system environment variable that specifies the timezone to be used by a YottaDB process, if it is not to use the default system timezone.

ydb_aio_nr_events

ydb_aio_nr_events (gtm_aio_nr_events): For Linux x86_64, the ydb_aio_nr_events environment variable controls the number of structures a process has per global directory to manage asynchronous writes, and therefore determines the number of concurrent writes a process can manage across all regions within a global directory. If not specified, the value controlled by ydb_aio_nr_events defaults to 128. If a process encounters a situation where it needs to perform an asynchronous write, but has no available slots with which to manage an additional one, it either falls back to synchronous writing if the write is blocking other actions, or defers the write until a slot becomes available as other writes complete. Linux allocates the structures on a system-wide basis with the setting of /proc/sys/fs/aio-max-nr. Therefore, you should configure this parameter to account for the needs (as determined by ydb_aio_nr_events or the default) of all processes using asynchronous I/O. When processes use multiple global directories with asynchronous I/O, their need for system resources increases accordingly. For example, if an environment runs 10,000 processes, each of which open two global directories and /proc/sys/fs/aio-max-nr is set to a value of 200,000 then ydb_aio_nr_events needs to be set to a value <= 200,000 / (10,000 * 2) = 10. Conversely if ydb_aio_nr_events is set to a value of 20, then aio-max-nr needs to be bumped up to (10,000 * 2 * 20) = 400,000. YottaDB captures the number of errors encountered when attempting to write database blocks for a region, and, barring problems with the storage subsystem, hitting an asynchronous write limit would constitute a primary (probably only) contribution to that value, which you can access with $$^%PEEKBYNAME("sgmnt_data.wcs_wterror_invoked_cntr",<region>).

ydb_app_ensures_isolation

When ydb_app_ensures_isolation is a comma-separated list of global variable names, those variables behave from process startup as if they had VIEW NOISOLATION set.

ydb_badchar

ydb_badchar (gtm_badchar) specifies the initial setting that determines whether YottaDB should raise an error when it encounters an illegal UTF-8 character sequence. This setting can be changed with a VIEW "[NO]BADCHAR" command, and is ignored for I/O processing and in M mode.

ydb_baktmpdir

ydb_baktmpdir (gtm_baktmpdir) specifies the directory where MUPIP BACKUP creates temporary files. If $ydb_baktmpdir is not defined, YottaDB uses the deprecated $GTM_BAKTMPDIR environment variable if defined, and otherwise the temporary files are created as follows:

  • MUPIP BACKUP DATABASE uses the directory of the backup destination for creating temporary files.

  • MUPIP BACKUP BYTESTREAM uses /tmp.

All processes performing updates during an online BACKUP must use the same directory and have write access to it.

ydb_boolean

ydb_boolean (gtm_boolean) specifies the initial setting that determines how YottaDB compiles Boolean expression evaluation (expressions evaluated as a logical TRUE or FALSE). If ydb_boolean is undefined or evaluates to an integer zero (0), YottaDB behaves as it would after a VIEW "NOFULL_BOOLEAN" and compiles such that it stops evaluating a Boolean expression as soon as it establishes a definitive result . Note that:

  • ydb_side_effects has an analogous impact on function argument evaluation order and implies "FULL_BOOLEAN" compilation, so VIEW "NOFULL_BOOLEAN" produces an error when ydb_side_effects is on.

  • If ydb_boolean evaluates to an integer one (1), YottaDB enables VIEW "FULL_BOOLEAN" compilation, which means that YottaDB ensures that within a Boolean expression, all side effect expression atoms, extrinsic functions ($$), external functions ($&), and $INCREMENT() execute in left-to-right order.

  • If ydb_boolean evaluates to an integer two (2), YottaDB enables VIEW "FULL_BOOLWARN" behavior, which means that YottaDB not only evaluates Boolean expressions like "FULL_BOOLEAN" but produces a BOOLSIDEFFECT warning when it encounters Boolean expressions that may induce side-effects; that is: expressions with side effects after the first Boolean operator - extrinsic functions, external calls, and $INCREMENT().

  • Boolean expressions without side effects will continue to be short-circuited whether or not ydb_boolean is 1 or 0. Error messages that could result if an expression were fully evaluated may not occur even with this setting enabled.

ydb_chset

ydb_chset (gtm_chset) determines the mode in which YottaDB compiles and operates. If it has a case-insensitive value of "UTF-8", YottaDB assumes that strings are encoded in UTF-8. In response to a value of "M" (or indeed anything other than "UTF-8"), YottaDB treats all 256 combinations of the 8 bits in a byte as a single character.

ydb_ci

ydb_ci (GTMCI) specifies the call-in table for function calls from C code to M code.

ydb_cm_<node-name>

ydb_cm_<node-name> is used by a GT.CM client process to locate the GT.CM server. <node-name> is an alphanumeric, which is used as a prefix for the GT.CM database segment file by the Global Directory of the client process. For detailed usage refer to the GT.CM Client section.

ydb_collate_n

ydb_collate_n (gtm_collate_n) specifies the shared library holding an alternative sequencing routine when using non-M standard (i.e., non-ASCII) collation. The syntax is ydb_collate_n=pathname where n is an integer from 1 to 255 that identifies the collation sequence, and pathname identifies the shared library containing the routines for that collation sequence.

ydb_compile

ydb_compile (gtmcompile) specifies the initial value of the $ZCOMPILE ISV. The SET command can alter the value of $ZCOMPILE in an active process.

ydb_coredump_filter

ydb_coredump_filter (gtm_coredump_filter) contains case-insensitive hexadecimal digits that sets the corresponding value to /proc/<pid>/coredump_filter (see man 5 core) at process startup without explicitly setting a value if unspecified. This controls the contents of core dumps generated by the process.

Note

Setting ydb_coredump_filter to -1 disables writing to /proc/<pid>/coredump_filter

ydb_crypt_config

ydb_crypt_config (gtmcrypt_config) specifies the location of the configuration file required for database encryption, Sequential file, PIPE, and FIFO device encryption and/or TLS support. A configuration file is divided into two sections: the database encryption section and the TLS section. The database encryption section contains a list of database files and their corresponding key files. You do not need to add a database encryption section if you are not using an encrypted database, or a TLS section if you are not using TLS for replication or sockets. The TLS section provides information needed for OpenSSL (in the reference plugin implementation) or other encryption package, such as the location of the root certification authority certificate in PEM format and leaf-level certificates with their corresponding private key files. Note that the use of the ydb_crypt_config environment variable requires prior installation of the libconfig package.

ydb_crypt_fips

ydb_crypt_fips (gtmcrypt_FIPS) specifies whether the plugin reference implementation should attempt to use either OpenSSL or Libgcrypt to provide database encryption that complies with FIPS 140-2. When the environment variable $ydb_crypt_fips is set to 1 (or evaluates to a non-zero integer or any case-independent string or leading substring of "TRUE" or "YES"), the plugin reference implementation attempts to use libgcrypt (from GnuPG) and libcrypto (OpenSSL) in "FIPS mode." Note that to comply with FIPS 140-2 you should be knowledgeable with that standard and take many steps beyond setting this environment variable. By default YottaDB does not enforce "FIPS mode.

ydb_crypt_plugin

ydb_crypt_plugin (gtm_crypt_plugin): If the environment variable ydb_crypt_plugin is defined and provides the path to a shared library relative to $ydb_dist/plugin, YottaDB uses $ydb_dist/plugin/$ydb_crypt_plugin as the shared library providing the plugin. If $ydb_crypt_plugin is not defined, YottaDB expects $ydb_dist/plugin/libgtmcrypt.so to be a symbolic link to a shared library providing the plugin. The expected name of the actual shared library is libgtmcrypt_cryptlib_CIPHER.so, for example, libgtmcrypt_openssl_AES256CFB.so.

ydb_custom_errors

ydb_custom_errors (gtm_custom_errors) specifies the complete path to the file that contains a list of errors that should automatically stop all updates on those region(s) of an instance which have the Instance Freeze mechanism enabled.

ydb_db_create_ver

ydb_db_create_ver (gtm_db_create_ver) directs MUPIP CREATE to create YottaDB r1.x (GT.M V6.x) compatible database file(s) when the environment variable is set to 6 or V6.

ydb_cur_gbldir

ydb_cur_gbldir specifies the current value of $ZGBLDIR set by the parent process using SET $ZGBLDIR. If ydb_cur_gbldir is not set it means the parent has not set a value to $ZGBLDIR, and is using the value set from the environment at process startup.

ydb_cur_gbldir was added to YottaDB effective release r1.36.

ydb_dbfilext_syslog_disable

ydb_dbfilext_syslog_disable (gtm_dbfilext_syslog_disable) Controls whether database file extensions are logged in the syslog or not. If the environment variable is set to a non-zero numeric value or case-independent string or leading substring of TRUE or YES, database file extensions are not logged to the syslog.

ydb_dbglvl

ydb_dbglvl (gtmdbglvl) specifies the YottaDB debug levels. The defined values can be added together to turn on multiple features at the same time. Note that the cumulative value specified in the logical or environment variable must currently be specified in decimal.

Note

Use of ydb_dbglvl is intended for debugging under the guidance of your YottaDB support channel. If you set ydb_dbglvl to a non-zero value, be aware that there will be a performance impact. We do not recommend its use in production.

Level

Value

Notes

GDL_None

0x00000000

No debugging

GDL_Simple

0x00000001

Regular assert checking, no special checks

GDL_SmStats

0x00000002

Print usage statistics at end of process

GDL_SmTrace

0x00000004

Trace each malloc/free (output to stderr)

GDL_SmDumpTrace

0x00000008

Dump malloc/free trace information on exit

GDL_SmAllocVerf

0x00000010

Perform verification of allocated storage chain for each call

GDL_SmFreeVerf

0x00000020

Perform simple verification of free storage chain for each call

GDL_SmBackfill

0x00000040

Backfill unused storage (cause exceptions if released storage is used)

GDL_SmChkAllocBackfill

0x00000080

Verify backfilled storage in GDL_AllocVerf while verifying each individual queue entry

GDL_SmChkFreeBackfill

0x00000100

Verify backfilled storage in GDL_FreeVerf while verifying each individual queue entry

GDL_SmStorHog

0x00000200

Each piece of storage allocated is allocated in an element twice the desired size to provide glorious amounts of backfill for overrun checking.

GDL_DumpOnStackOFlow

0x00000400

When get a stack overflow or out-of-memory error, generate a core

GDL_ZSHOWDumpOnSignal

0x00000800

Don't supress YDB_FATAL file creation when get a signal

GDL_PrintIndCacheStats

0x00001000

Print indirect cacheing stats

GDL_PrintCacheStats

0x00002000

Print stats on $Piece and UTF8 cacheing (debug only)

GDL_DebugCompiler

0x00004000

Turn on compiler debugging

GDL_SmDump

0x00008000

Do full blown storage dump – only useful in debug mode

GDL_PrintEntryPoints

0x00010000

Print address of entry points when they are loaded/resolved

GDL_PrintSockIntStats

0x00020000

Print Socket interrupt stats on exit

GDL_SmInitAlloc

0x00040000

Initialize all storage allocated or deallocated with 0xdeadbeef

GDL_PrintPipeIntStats

0x00080000

Print Pipe/Fifo(rm) interrupt stats on exit

GDL_IgnoreAvailSpace

0x00100000

Allow gdsfilext/mu_cre_file (UNIX) to ignore available space

GDL_PrintPMAPStats

0x00200000

Print process memory map on exit (using pmap or procmap utility)

GDL_AllowLargeMemcpy

0x00400000

Bypass the 1GB sanity check in gtm_memcpy_validate_and_execute()

define GDL_UseSystemMalloc

0x80000000

Use the system's malloc(), disabling all the above GDL_Sm options

ydb_db_startup_max_wait

ydb_db_startup_max_wait (gtm_db_startup_max_wait) specifies how long processes should wait for a resolution of any resource conflict when they first access a database file or the replication instance file. YottaDB uses semaphores maintained using UNIX Inter-Process Communication (IPC) services to ensure orderly initialization and shutdown of database files, replication instance files and associated shared memory. Normally, the IPC resources are held in an exclusive state only for very brief intervals. However, under unusual circumstances that might include extremely large numbers of simultaneous database or instance file initializations, a long-running MUPIP operation involving standalone access (like INTEG -FILE or RESTORE), an OS overload or an unpredicted process failure, the resources might remain unavailable for an unanticipated length of time. $ydb_db_startup_max_wait specifies how long to wait for the resources to become available:

  • -1 - Indefinite wait until the resource becomes available; the waiting process uses the ydb_procstuckexec mechanism at approximately 48 and 96 seconds.

  • 0 - No wait - if the resource is not immediately available, give a DBFILERR error with an associated SEMWT2LONG

  • > 0 - Seconds to wait - rounded to the nearest multiple of eight (8); if the specification is 96 or more seconds, the waiting process uses the ydb_procstuckexec mechanism at one half the wait and at the end of the wait; if the resource remains unavailable, the process issues DBFILERR error with an associated SEMWT2LONG

ydb_dist

ydb_dist (gtm_dist) specifies the path to the directory containing the YottaDB system distribution. ydb_dist must be defined for each user. If you are not using the ydb script or sourcing ydb_env_set, consider defining ydb_dist in the login file or as part of the default system environment.

Effective release r1.30, at process initialization YottaDB ensures that gtm_dist is set to $ydb_dist.

ydb_dmterm

ydb_dmterm (gtm_dmterm) specifies an initial value at process startup for VIEW "DMTERM". A case-insensitive value of "1", "yes", or "true" establishes a DMTERM state at process initiation where direct mode uses default terminal characteristics and ignores application settings for $PRINCIPAL; all other values, including no value, result in the default VIEW "NODMTERM" behavior.

ydb_dollartest

ydb_dollartest provides an initial value for $TEST. When ydb_dollartest is set to 0 or 1, the value of $TEST will be set to 0 or 1 respectively. If ydb_dollartest is undefined then the value of $TEST will be set to 1.

ydb_env_translate

ydb_env_translate (gtm_env_translate) specifies the path to a shared library to implement the optional YottaDB environment translation facility to aid application portability across platforms by translating strings into global directory references.

ydb_error_on_jnl_file_lost

ydb_error_on_jnl_file_lost (gtm_error_on_jnl_file_lost) causes a runtime error when set to 1 in case of problems with journaling (disk space issues etc.). Setting this environment variable to 0 (or having it undefined) is the default behavior which is to turn off journaling in case of problems.

ydb_etrap

ydb_etrap (gtm_etrap) specifies an initial value of $ETRAP to override the default value of "B" for $ZTRAP as the base level error handler. The ydb_env_set script sets ydb_etrap to "Write:(0=$STACK) ""Error occurred: "",$ZStatus,!".

ydb_extract_nocol

ydb_extract_nocol (gtm_extract_nocol) specifies whether a MUPIP JOURNAL EXTRACT (when used without RECOVER or ROLLBACK) on the journal file of a database region with custom collation should use the default collation if it is not able to read the database file. In a situation where the database file is inaccessible or the replication instance is frozen with a critical section required for the access held by another process and the environment variable ydb_extract_nocol is defined and evaluates to a non-zero integer, MUPIP JOURNAL EXTRACT issues the DBCOLLREQ warning and proceeds with the extract using the default collation. If ydb_extract_nocol is not set or evaluates to a value other than a positive integer, MUPIP JOURNAL EXTRACT exits with the SETEXTRENV error.

Note

If default collation is used for a database with custom collation, the subscripts reported by MUPIP JOURNAL -EXTRACT are those stored in the database, which may differ from those used by application logic.

ydb_flushoncallout

ydb_flushoncallout specifies whether the process should startup with VIEW FLUSHONCALLOUT. If set to a non-zero numeric value. "yes" or "TRUE" (case-insensitive), or a leading substring thereof, causes the process to startup with VIEW FLUSHONCALLOUT. Any other value, or no value causes the process to startup with VIEW NOFLUSHONCALLOUT.

ydb_flushoncallout was added to YottaDB effective release r1.36.

ydb_fullblockwrites

ydb_fullblockwrites (gtm_fullblockwrites) specifies whether a YottaDB process should write a full filesystem, or full database block, worth of bytes when writing a database block that is not full. Depending on your IO subsystem, writing a full block worth of bytes (even when there are unused garbage bytes at the end) may result in better database IO performance by replacing a low level read-modify-read IO operation with a single write operation.

ydb_fullblockwrites is deprecated in YottaDB effective release r1.36 and no longer maintained or tested.

ydb_gbldir

ydb_gbldir (gtmgbldir) specifies the initial value of $ZGBLDIR, which identifies the global directory, required to access M global variables.

ydb_gbldir_translate

ydb_gbldir_translate provides the path to a shared library to allow a set of $ZGBLDIR to be transformed for application portability across platforms, using the optional YottaDB global directory translation facility. This is similar to the the optional YottaDB environment translation facility provided by ydb_env_translate above. ydb_gbldir_translate was added effective release r1.30.

ydb_gdscert

ydb_gdscert (gtm_gdscert) specifies the initial setting that controls whether YottaDB processes should test updated database blocks for structural damage. If it is defined, and evaluates to a non-zero integer or any case-independent string or leading substring of "TRUE" or "YES", YottaDB performs a block-level integrity check on every block as a process commits it. Within a running process, VIEW "GDSCERT":value controls this setting. By default, YottaDB does not check database blocks for structural damage, because the impact on performance is usually unwarranted.

ydb_hostname

If the environment variable ydb_hostname is set, YottaDB uses its value instead of the actual hostname, to record in each database file header as to which host has that file open. This allows multiple Kubernetes pods which have random hostnames, but share IPC resources, to access database files across pods. Note that pods accessing a database file with the same ydb_hostname, but without sharing IPC resources, will result in structural damage to that database file. So use this feature with caution.

ydb_hugetlb_shm

ydb_hugetlb_shm (gtm_hugetlb_shm) specifies the initial value that determines whether a YottaDB process should use hugepages to back shared memory. If it is defined, and evaluates to a non-zero integer or any case-independent string or leading substring of "TRUE" or "YES" in a YottaDB process creating shared memory, YottaDB attempts to back all such shared memory segments with hugepages, using the default hugepage size. If hugepages cannot be used, YottaDB backs the shared memory with base pages instead, and attempts to pin the shared memory if requested with $ydb_pinshm.

ydb_hupenable

ydb_hupenable (gtm_hupenable) specifies the initial value that determines whether a YottaDB process should recognize a disconnect signal from a PRINCIPAL device that is a terminal. If it is defined, and evaluates to a non-zero integer or any case-independent string or leading substring of "TRUE" or "YES", the process receives a TERMHANGUP error if the OS signals that the terminal assigned to the process as the PRINCIPAL device has disconnected. Within a running process, USE $PRINCIPAL:[NO]HUP[ENABLE] controls this behavior. By default, YottaDB ignores such a signal, but a process that ignores the signal may subsequently receive an IOEOF or a TERMWRITE error from an attempt to respectively READ from, or WRITE to the missing device. YottaDB terminates a process that ignores more than one of these messages and, if the process is not in Direct Mode, sends a NOPRINCIO message to the syslog.

ydb_hupenable was added to YottaDB effective release r1.34.

ydb_icu_version

ydb_icu_version (gtm_icu_version) specifies the MAJOR VERSION and MINOR VERSION numbers of the desired ICU. For example "3.6" denotes ICU-3.6. If $ydb_chset has the value "UTF-8", YottaDB requires libicu with version 3.6 or higher. If you must chose between multiple versions of libicu or if libicu has been compiled with symbol renaming enabled, YottaDB requires ydb_icu_version to be explicitly set. Please see the section on Configuring and operating YottaDB with Unicode™ support (optional) for M for more information.

ydb_ipv4_only

ydb_ipv4_only (gtm_ipv4_only) specifies whether a Source Server should establish only IPv4 connections with a Receiver Server or sockets associated with a SOCKET device. If it is defined, and evaluates to a non-zero integer, or any case-independent string or leading substring of "TRUE" or "YES", the Source Server establishes only IPv4 connections with the Receiver Server. ydb_ipv4_only is useful for environments where different server names are not used for IPv4 and IPv6 addresses.

ydb_jnl_release_timeout

ydb_jnl_release_timeout (gtm_jnl_release_timeout) specifies the number of seconds that a replicating Source Server waits when there is no activity on an open journal file before closing it. The default wait period is 300 seconds (5 minutes). If $ydb_jnl_release_timeout specifies 0, the Source Server keeps the current journal files open until shutdown. The maximum value for $ydb_jnl_release_timeout is 2147483 seconds.

ydb_keep_obj

ydb_keep_obj (gtm_keep_obj) specifies whether the ydbinstall script should delete the object files from the YottaDB installation directory. If ydb_keep_obj is set to "Y", the ydbinstall script leaves object files; by default, ydbinstall deletes object files after archiving them in a shared library.

ydb_lct_stdnull

ydb_lct_stdnull (gtm_lct_stdnull) specifies whether a YottaDB process should use standard collation for local variables with null subscripts or historical null collation.

ydb_linktmpdir

ydb_linktmpdir (gtm_linktmpdir) identifies a directory (defaulting to $ydb_tmp, which in turn defaults to /tmp, if unspecified) where YottaDB creates a small control file (Relinkctl), for each auto-relink enabled directory which a YottaDB process accesses while searching through $ZROUTINES. The names of these files are of the form ydb-relinkctl-<murmur> where <murmur> is a hash of the realpath() to an auto-relink directory; for example: /tmp/ydb-relinkctl-f0938d18ab001a7ef09c2bfba946f002. With each Relinkctl file, YottaDB creates and associates a block of shared memory that contains associated control structures. Among the structures is a cycle number corresponding to each routine found in the routine directory; a change in the cycle number informs a process that it may need to determine whether there is a new version of a routine. Although YottaDB only creates relinkctl records for routines that actually exist on disk, it may increment cycle numbers for existing relinkctl records even if they no longer exist on disk.

ydb_locale

ydb_locale (gtm_locale) specifies a locale to use (LC_CTYPE would be set to this value) if the ydb_chset environment variable is set to UTF-8. If not set, the current value of LC_CTYPE is used. This environment variable is ignored if ydb_chset is not set to UTF-8.

ydb_local_collate

ydb_local_collate (gtm_local_collate) specifies an alternative collation sequence for local variables.

ydb_log

ydb_log (gtm_log) specifies a directory where the gtm_secshr_log file is stored. The gtm_secshr_log file stores information gathered in the gtmsecshr process. YottaDB recommends that a system-wide default be established for ydb_log so that gtmsecshr always logs its information in the same directory, regardless of which user's YottaDB process invokes gtmsecshr. In conformance with the Filesystem Hierarchy Standard, YottaDB recommends /var/log/yottadb/$ydb_rel as the value for $ydb_log unless you are installing the same version of YottaDB in multiple directories. Note that $ydb_rel can be in the form of the current YottaDB release and platform. If you do not set $ydb_log, YottaDB creates log files in a directory in /tmp. However, this is not recommended because it makes YottaDB log files vulnerable to the retention policy of a temporary directory.

Note

In current versions, gtmsecshr logs its messages in the system log and the environment variable ydb_log is ignored.

ydb_lvnullsubs

ydb_lvnullsubs (gtm_lvnullsubs) specifies the initialization of VIEW [NEVER][NO]LVNULLSUBS at process startup. The value of the environment variable can be 0 which is equivalent to VIEW “NOLVNULLSUBS”, 1 (the default) which is equivalent to VIEW “LVNULLSUBS” or 2, which is equivalent to VIEW “NEVERLVNULLSUBS”.

ydb_malloc_limit

ydb_malloc_limit (gtm_malloc_limit) specifies the initial value of $ZMALLOCLIM at process startup. An integer value specifies a number of bytes of process memory, which, if exceeded, cause YottaDB to issue a MALLOCCRIT error. The default is 0, which indicates no warning limit on process memory utilization. When the environment variable specifies a positive value, YottaDB uses the smaller of that value (with a minimum of 2.5MB) and any OS defined amount for the value of $ZMALLOCLIM. A value of minus one (-1) provides a value of half the system imposed limit if any. YottaDB does not give errors or messages about its choice for $ZMALLOCLIM between a specified value and some other more appropriate value, so if the application needs to verify the result, it should examine the resulting ISV value.

ydb_maxtptime

ydb_maxtptime (gtm_zmaxtptime) specifies the initial value of the $ZMAXTPTIME Intrinsic Special Variable, which controls whether and when YottaDB issues a TPTIMEOUT error for a TP transaction that runs too long. ydb_maxtptime specifies time in seconds and the default is 0, which indicates "no timeout" (unlimited time). The maximum value of ydb_maxtptime is 60 seconds and the minimum is 0; YottaDB ignores ydb_maxtptime if it contains a value outside of this recognized range. This range check does not apply to SET $ZMAXTPTIME.

ydb_max_indrcache_count/ydb_max_indrcache_memory

ydb_max_indrcache_count (gtm_max_indrcache_count) and ydb_max_indrcache_memory (gtm_max_indrcache_memory) control the cache of compiled code for indirection/execute. ydb_max_indrcache_count is the maximum number of entries in the cache (defaulting to 128) and ydb_max_indrcache_memory is maximum memory (in KiB, defaulting to 128). When the number of cache entries exceeds $ydb_max_indrcache_count, or the memory exceeds $ydb_max_indrcache_memory KiB, YottaDB discards the entire cache and starts over.

ydb_max_sockets

ydb_max_sockets (gtm_max_sockets) specifies the maximum number of client connections for socket devices. The default is 64. While it must be large enough to accommodate the actual need, each reservation requires some memory in socket structures, so setting this number unnecessarily high causes requires a bit of additional memory for no benefit.

ydb_max_storalloc

ydb_max_storalloc (gtm_max_storalloc) limits the amount of memory (units in bytes) a YottaDB process is allowed to allocate before issuing a MEMORY (and MALLOCMAXUNIX) error. This helps in tracking memory allocation issues in the application.

ydb_memory_reserve

ydb_memory_reserve (gtm_memory_reserve) specifies the size in kilobytes of the reserve memory that YottaDB should use in handling and reporting an out-of-memory condition. The default is 64 (KiB). Setting this too low can impede investigations of memory issues, but YottaDB only uses this reserve when a process runs out of memory so it almost never requires actual memory, only address space.

ydb_msgprefix

ydb_msgprefix specifies a prefix for YottaDB messages generated by a process, with the prefix defaulting to "YDB", e.g., YDB-I-DBFILEXT. Previously, the prefix was always "GTM". A value of "GTM" retains the previous format.

ydb_mstack_crit_threshold

ydb_mstack_crit_threshold (gtm_mstack_crit_threshold) specifies an integer between 15 and 95 defining the percentage of the stack which should be used before YottaDB emits a STACKCRIT warning. If the value is below the minimum or above the maximum, YottaDB uses the minimum or maximum respectively. The default is 90.

ydb_mstack_size

ydb_mstack_size (gtm_stack_size) specifies the M stack size (in KiB). If ydb_mstack_size is not set or set to 0, YottaDB uses the default M stack size (that is, 272KiB). The minimum supported size is 25 KiB; YottaDB reverts values smaller than this to 25 KiB. The maximum supported size is 10000 KiB; YottaDB reverts values larger than this to 10000 KiB.

ydb_mupjnl_parallel

ydb_mupjnl_parallel (gtm_mupjnl_parallel) defines the number of processes or threads used by MUPIP JOURNAL RECOVER/ROLLBACK when the invoking command does not have a -PARALLEL qualifier. When defined with no value, it specifies one process or thread per region. When undefined or defined to one (1), it specifies MUPIP should process all regions without using additional processes or threads. When defined with an integer value greater than one (1), it specifies the maximum number of processes or threads for MUPIP to use. If the value is greater than the number of regions, MUPIP never uses more processes or threads than there are regions. If it is less than the number of regions, MUPIP allocates work to the additional processes or threads based on the time stamps in the journal files.

ydb_nocenable

ydb_nocenable (gtm_nocenable) specifies whether the $principal terminal device should ignore <CTRL-C> or use <CTRL-C> as a signal to place the process into direct mode; a USE command can modify this device characteristic. If ydb_nocenable is defined and evaluates to a non-zero integer or any case-independent string or leading substring of "TRUE" or "YES", $principal ignores <CTRL-C>. If ydb_nocenable is not set or evaluates to a value other than a positive integer or any case-independent string or leading substring of "FALSE" or "NO", <CTRL-C> on $principal places the process into direct mode at the next opportunity (usually at a point corresponding to the beginning of the next source line).

ydb_nofflf

ydb_nofflf (gtm_nofflf) specifies the default WRITE # behavior for STREAM and VARIABLE format sequential files. If it is set to 1, Y[ES] or T[RUE], WRITE # writes only a form-feed <FF> character in conformance to the M standard. If it is not defined or set to 0, N[O] or F[ALSE], WRITE # writes <FF><LF> characters. The [NO]FFLF deviceparameter for USE and OPEN commands takes precedence over any value of ydb_nofflf.

ydb_nofflf was added to YottaDB effective release r1.34.

ydb_non_blocked_write_retries

ydb_non_blocked_write_retries (gtm_non_blocked_write_retries) modifies WRITE behavior for FIFO, PIPE, or non-blocking sockets. A WRITE which would block is retried up to the number specified with a 100 milliseconds delay between each retry. The default value is 10 times. If all retries block, the WRITE command issues a %SYSTEM-E-ENO11 (EAGAIN) error. For more details, refer to PIPE Device Examples in the Programmers Guide.

ydb_nontprestart_log_delta

ydb_nontprestart_log_delta (gtm_nontprestart_log_delta) specifies the frequency with which YottaDB reports non-transaction restarts to the syslog. A value of 1 means that every non-transaction restart is to be reported. If ydb_nontprestart_log_delta is not defined, YottaDB initializes ydb_nontprestart_log_delta to 0, meaning that no restarts are to be reported, regardless of the value of ydb_nontprestart_log_first.

ydb_nontprestart_log_first

ydb_nontprestart_log_first (gtm_nontprestart_log_first) specifies the initial number of non-transaction restarts which YottaDB should report before pacing subsequent non-transaction restart reports to the syslog using the ydb_nontprestart_log_delta value. If ydb_nontprestart_log_delta is defined and ydb_nontprestart_log_first is not defined, YottaDB initializes ydb_nontprestart_log_first to 0.

ydb_noundef

ydb_noundef (gtm_noundef) specifies the initial setting that controls whether a YottaDB process should treat undefined global or local variables as having an implicit value of an empty string. If it is defined, and evaluates to a non-zero integer or any case-independent string or leading substring of "TRUE" or "YES", then YottaDB treats undefined variables as having an implicit value of an empty string. The VIEW "[NO]UNDEF" command can alter this behavior in an active process. By default, YottaDB signals an error on an attempt to use the value of an undefined variable.

ydb_obfuscation_key

ydb_obfuscation_key (gtm_obfuscation_key) : If $ydb_obfuscation_key specifies the name of the file readable by the process, the encryption reference plug-in uses a cryptographic hash of the file's contents as the XOR mask for the obfuscated password in the environment variable ydb_passwd. When ydb_obfuscation_key does not point to a readable file, the plugin computes a cryptographic hash using a mask based on the value of $USER and the inode of the yottadb executable to use as a mask. $ydb_passwd set with a $ydb_obfuscation_key allows access to all users who have the same $ydb_obfuscation_key defined in their environments. However, $ydb_passwd set without $ydb_obfuscation_key can be used only by the same $USER using the same YottaDB distribution.

ydb_passwd

ydb_passwd (gtm_passwd) specifies the obfuscated (not encrypted) password of the GNU Privacy Guard key ring. When the environment variable $ydb_passwd is set to "", YottaDB invokes the default GTMCRYPT passphrase prompt defined in the reference implementation of the plugin to obtain a passphrase at process startup and uses that value as $ydb_passwd for the duration of the process.

ydb_patnumeric

ydb_patnumeric (gtm_patnumeric) specifies the value of the read-only ISV $ZPATNUMERIC that determines how YottaDB interprets the patcode "N" used in the pattern match operator. The SET command can alter the value of $ZPATNUMERIC in an active process.

ydb_pattern_file/ydb_pattern_table

ydb_pattern_file (gtm_pattern_file) and ydb_pattern_table (gtm_pattern_table) specify alternative patterns for the pattern (?) syntax. Refer to the Internationalization chapter in the Programmer's Guide for additional information.

ydb_pinshm

ydb_pinshm (gtm_pinshm) specifies whether a YottaDB process should attempt to pin shared memory it creates into physical memory. If it is defined and evaluates to a non-zero integer or any case-independent string or leading substring of "TRUE" or "YES" in a process creating shared memory, YottaDB attempts to pin shared memory used for database global buffers, replication buffers, and routine buffers into physical memory. As hugepages are implicitly locked in physical memory, YottaDB does not attempt to pin shared memory buffers backed by hugepages. $ydb_pinshm does not pin memory used by online INTEG (integ snapshot). Pinning may not succeed due to insufficient physical memory and/or OS configuration. By default, YottaDB does not attempt to pin shared memory.

ydb_poollimit

ydb_poollimit (gtm_poollimit) restricts the number of global buffers a process uses in order to limit the potential impact on other processes. It is intended for use by MUPIP REORG, since it has the potential to "churn" global buffers; the value is of the form n[%]. When it ends with a per-cent sign (%), the number is taken as a percentage of the configured global buffers and otherwise as an ordinal number of preferred buffers; standard M parsing and integer conversions apply. Note that this environment variable applies to all regions accessed by a process; the VIEW command for this feature allows finer grained control. MUPIP REORG uses this facility to limit its buffers with a default of 64 if ydb_poollimit is not specified. Note that this may slightly slow a standalone REORG but can be overridden by defining ydb_poollimit as 0 or "100%".

ydb_principal

ydb_principal (gtm_principal) specifies the value for $PRINCIPAL, which designates an alternative name (synonym) for the principal $IO device.

ydb_principal_editing

ydb_principal_editing (gtm_principal_editing) specifies the initial settings for $PRINCIPAL for the following colon-delimited deviceparameters: [NO]EDITING, [NO]EMPTERM and [NO]INSERT; in an active process the USE command can modify these device characteristics.

Note

The YottaDB direct mode commands have a more extensive capability in this regard, independent of the value of this environment variable.

ydb_procstuckexec

ydb_procstuckexec (gtm_procstuckexec) specifies a shell command or a script to execute when any of the following conditions occur:

  • A one minute wait on a region due to an explicit MUPIP FREEZE or an implicit freeze, such as BACKUP, INTEG -ONLINE, and so on.

  • MUPIP actions find kill_in_prog (KILLs in progress) to be non-zero after a one minute wait on a region. Note that YottaDB internally maintains a list of PIDs (up to a maximum of 8 PIDs) currently doing a KILL operation.

  • A process encounters conditions that produce the following syslog messages: BUFOWNERSTUCK, INTERLOCK_FAIL, JNLPROCSTUCK, SHUTDOWN, WRITERSTUCK, MAXJNLQIOLOCKWAIT, MUTEXLCKALERT, SEMWT2LONG, and COMMITWAITPID.

You can use this as a monitoring facility for processes holding a resource for an unexpected amount of time. Typically, for the shell script or command pointed to by ydb_procstuckexec, you would write corrective actions or obtain the stack trace of the troublesome processes (using their PIDs). YottaDB passes arguments to the shell command/script in the order specified as follows:

  • condition is the name of the condition. For example, BUFOWNERSTUCK, INTERLOCK_FAIL, and so on.

  • waiting_pid is the PID of the process reporting the condition.

  • blocking_pid is the PID of the process holding a resource.

  • count is the number of times the script has been invoked for the current condition (1 for the first occurrence).

Each invocation generates an syslog message and if the invocation fails, an error message to the syslog. The shell script should start with a line beginning with #! that designates the shell.

Instead of creating your own custom script, we recommend that you use the %YDBPROCSTUCEXEC utility program included with YottaDB. Set $ydb_procstuckexec / $gtm_procstuckexec to "$ydb_dist/yottadb -run %YDBPROCSTUCKEXEC" to use this standard utility program. In this case, ensure that all processes have the same value for $ydb_tmp / $gtm_tmp.

Note

Make sure that user processes have sufficient space and permissions to run the shell command/script. For example - for the script to invoke the debugger, the process must be of the same group or have a way to elevate privileges.

ydb_prompt

ydb_prompt (gtm_prompt) specifies the initial value of the ISV $ZPROMPT, which controls the YottaDB direct mode prompt. The SET command can alter the value of $ZPROMPT in an active process. By default, the direct mode prompt is "YDB>".

ydb_quiet_halt

ydb_quiet_halt (gtm_quiet_halt) specifies whether YottaDB should disable the FORCEDHALT message when the process is stopped via MUPIP STOP or by a SIGTERM signal (as sent by some web servers).

ydb_readline

ydb_readline when set to 1 (recommended) or a case insensitive t[rue] or y[es], specifies that M Direct Mode, LKE, DSE and MUPIP should use GNU Readline if it is installed on the system. Otherwise, YottaDB direct mode uses a traditional implementation that is part of YottaDB, whereas LKE, DSE and MUPIP have no recall capability, and very basic line editing. A short summary of YottaDB use of Readline is provided here; refer to the Readline documentation for details. ydb_env_set sets ydb_readline to 1.

Command history is saved in $HOME/.ydb_{YottaDB,DSE,LKE,MUPIP}_history. When Readline is enabled:

  • Direct mode RECALL displays Readline history when entered at the beginning of the line.

  • Recalling and editing prior commands with ^ or ! work when they are entered at the beginning of the line. Ending the line with :p prints the recalled and edited command, instead of executing it.

  • Recalling and editing commands also executes them, unlike the RECALL command implemented by YottaDB.

  • Entering UTF-8 mode characters works in M mode.

  • Settings are read from $HOME/.inputrc, whose location can be overridden by the INPUTRC environment variable. The application name for use in $if statements in the settings file is YottaDB.

  • If history-size is not set, it defaults to 1,000. The history file on disk is always limited to 1,000 entries, no matter the setting of history-size.

  • Signals are handled by YottaDB and not by Readline.

Examples of history expansion:

  • !!: Recall last command

  • !$: Last argument of last command

  • !nnn: Execute line in history number nnn

  • !nnn:p: Print line (but don't execute) nnn, and add it to the history to the end. You can press up arrow to recall that command for editing.

  • !?xxxx: Execute line containing text xxxx. BE CAREFUL WITH THIS ONE. It can lead to unexpected items getting executed.

  • ^string1^string2^: In the last command, replace string1 with string2, and execute it.

  • !nnn:s/old/new/: In history item nnn, replace old with new and execute it.

Limitations include:

  • DSE/LKE/MUPIP

    • There is no history listing (equivalent to the direct mode RECALL command).

    • History expansion module works only in direct mode.

  • Direct mode

    • Only traditional characters terminate input lines (CR, LF, FF, and their UTF-8 variants); alternate terminators are not supported. (YottaDB direct mode has the ability to terminate input using the TERMINATOR deviceparameter.)

    • Wrap on device width (set using WIDTH deviceparameter) is not supported.

    • MUPIP INTRPT (SIGUSR1) turns off line editing on the line being entered. You can still enter more characters or cancel the line using CTRL-C.

  • Readline is not supported for the READ command.

ydb_recompile_newer_src

ydb_recompile_newer_src when set to 1, t[rue], or y[es], specifies that a ZLINK/DO/GOTO/ZBREAK/ZGOTO/ZPRINT/$TEXT should recompile the .m file only if it has a newer modification time than the corresponding .o file. The default behavior is for the .m file to be recompiled if its modification time is later than OR equal to that of the corresponding .o file. ydb_recompile_newer_src was added effective release r1.30.

ydb_rel

ydb_rel (gtmversion) (not used by YottaDB directly) - The current YottaDB version. The ydb_env_set script uses $ydb_rel to set other environment variables.

ydb_repl_filter_timeout

ydb_repl_filter_timeout (gtm_repl_filter_timeout) can be set to an integer value indicating the timeout (in seconds) that the replication source server sets for a response from the external filter program. A value less than 32 would be treated as if 32 was specified. A value greater than 131072 (2**17) would be treated as if 131072 was specified. The default value of the timeout (if env var is not specified) is 64 seconds. This provides the user a way to avoid seeing FILTERTIMEDOUT errors from the source server on relatively slower systems.

ydb_repl_instance

ydb_repl_instance (gtm_repl_instance) specifies the location of the replication instance file when database replication is in use.

ydb_repl_instname

ydb_repl_instname (gtm_repl_instname) specifies a replication instance name that uniquely identifies an instance. The replication instance name is immutable. The maximum length of a replication instance name is 15 bytes. Note that the instance name is not the same as the name of the replication instance file (ydb_repl_instance). You need to specify a replication instance name at the time of creating a replication instance file. If you do not define ydb_repl_instname, you need to specify an instance name using -NAME=<instance_name> with MUPIP REPLICATE -INSTANCE_CREATE.

ydb_repl_instsecondary

ydb_repl_instsecondary (gtm_repl_instsecondary) specifies the name of the replicating instance in the current environment. YottaDB uses $ydb_repl_instsecondary if the -instsecondary qualifer is not specified.

ydb_retention

ydb_retention (gtm_retention) (not used by YottaDB directly) - Journal files and temporary files older than the number of days specified by ydb_retention (gtm_retention if not specified; defaulting to 42 days), are deleted by sourcing the ydb_env_set file, which can be invoked explicitly, or as part of executing the ydb script.

ydb_routines

ydb_routines (gtmroutines) specifies the initial value of the $ZROutines ISV, which specifies where to find object and source code. The SET command can alter the value of $ZROUTINES in an active process.

ydb_side_effects

ydb_side_effects (gtm_side_effects): When the environment variable ydb_side_effects is set to one (1) at process startup, YottaDB generates code that performs left to right evaluation of actual list arguments, function arguments, operands for non-Boolean binary operators, SET arguments where the target destination is an indirect subscripted glvn, and variable subscripts. When the environment variable is not set or set to zero (0), YottaDB retains its traditional behavior, which re-orders the evaluation of operands using rules intended to improve computational efficiency. This reordering assumes that functions have no side effects, and may generate unexpected behavior (x+$increment(x) is a pathological example). When ydb_side_effects is set to two (2), YottaDB generates code with the left-to-right behavior, and also generates SIDEEFFECTEVAL warning messages for each construct that potentially generates different results depending on the order of evaluation. As extrinsic functions and external calls are opaque to the compiler at the point of their invocation, it cannot statically determine whether there is a real interaction. Therefore, SIDEEFFECTEVAL warnings may be much more frequent than actual side effect interactions and the warning mode may be most useful as a diagnostic tool to investigate problematic or unexpected behavior in targeted code rather than for an audit of an entire application. Note that a string of concatenations in the same expression may generate more warnings than the code warrants. Other values of the environment variable are reserved for potential future use by YottaDB. It is important to note that ydb_side_effects affects the generated code, and must be in effect when code is compiled - the value when that compiled code is executed is irrelevant. Note also that XECUTE and auto-ZLINK, explicit ZLINK and ZCOMPILE all perform run-time compilation subject to the characteristics selected when the process started. Please be aware it is an unsafe programming practice when one term of an expression changes a prior term in the same expression. The environment variable ydb_boolean may separately control short-circuit evaluation of Boolean expressions but a setting of 1 (or 2) for ydb_side_effects causes the same boolean evaluations as setting ydb_boolean to 1 (or 2). Note that warning reports for the two features are separately controlled by setting their values to 2. The differences in the compilation modes may include not only differences in results, but differences in flow of control when the code relies on side effect behavior.

ydb_snaptmpdir

ydb_snaptmpdir (gtm_snaptmpdir) specifies the location to place the temporary "snapshot" file created by facilities such as MUPIP INTEG ONLINE. If $ydb_snaptmpdir is not defined, YottaDB uses the deprecated $GTM_BAKTMPDIR environment variable if defined, and otherwise uses the current working directory. All processes performing updates during an online INTEG must use the same directory and have write access to it.

ydb_string_pool_limit

ydb_string_pool_limit (gtm_string_pool_limit) is used for the initial value of $ZSTRPLLIM, when it specifies a positive value.

ydb_statsdir

ydb_statsdir (gtm_statsdir) specifies the directory for database files into which processes that have opted-in to sharing global statistics place their statistics as binary data. If you do not explicitly define this environment variable for a process, YottaDB defines this to the evaluation of $ydb_tmp, which defaults to /tmp. All processes that share statistics MUST use the same value for $ydb_statsdir. YottaDB suggests that you point ydb_statsdir at a tmpfs or ramfs. These database files have a name derived from the user defined database file name and a .gst extension. They are not usable as normal database files by application code, except to read statistics. YottaDB automatically creates and deletes these database files as needed. Under normal operation, applications do not need to manage them explicitly. The mapping of ^%YGS to statistics database files is managed by YottaDB transparently to applications with global directories. The ^%YGBLSTAT utility program gathers and reports statistics from nodes of ^%YGS(region,pid).

ydb_statshare

ydb_statshare (gtm_statshare) specifies an initial value for the VIEW "[NO]STATSHARE" setting.

ydb_stdxkill

ydb_stdxkill (gtm_stdxkill) enables the standard-compliant behavior to kill local variables in the exclusion list if they had an alias that was not in the exclusion list. By default, this behavior is disabled.

ydb_sysid

ydb_sysid (gtm_sysid) specifies the value for the second piece of the $SYSTEM intrinsic special variable.

ydb_tls_passwd_<label>

ydb_tls_passwd_<label> (gtmtls_passwd_<label>) specifies the obfuscated password of the encrypted private key pair. You can obfuscate passwords using the 'maskpass' utility provided along with the encryption plugin. If you choose to use unencrypted private keys, set the ydb_tls_passwd_<label> environment variable to a non-null dummy value; this prevents inappropriate prompting for a password.

ydb_tmp

ydb_tmp (gtm_tmp) specifies a directory where socket files used for communication between gtmsecshr and YottaDB processes are stored. All processes using the same YottaDB installation (i.e., the same ydb_dist) should have the same $ydb_tmp. YottaDB recommends setting ydb_tmp to a location:

  • which is unique to a YottaDB version, e.g., /tmp/r1.34_x86_64; and

  • where operation procedures and policies can ensure that the contents are not removed when there are active YottaDB processes.

If appropriate, set ydb_tmp to a directory location in tmpfs or ramfs on Linux.

If ydb_tmp is not defined, YottaDB uses the /tmp directory which may disrupt active gtmsecshr operations when /tmp is either cleared manually or by the retention policies of the operating system.

ydb_tpnotacidtime

ydb_tpnotacidtime (gtm_tpnotacidtime) specifies the maximum time that a YottaDB process waits for a non-isolated timed command (HANG, JOB, LOCK, OPEN, READ, WRITE /* or ZALLOCATE) running within a transaction to complete before it releases all critical sections it owns and sends a TPNOTACID information message to the system log. A YottaDB process owns critical sections on all or some of the regions participating in a transaction, only during final retry attempts (when $TRESTART>2). ydb_tpnotacidtime specifies time in seconds to millisecond precision (three decimal places); the default is 2 seconds. The maximum value of ydb_tpnotacidtime is 30 and the minimum is 0. If ydb_tpnotacidtime specifies a time outside of this range, YottaDB uses the default value. YottaDB releases critical sections in a final retry attempt to provide protection from certain risky coding patterns which, because they are not isolated, can cause deadlocks (in the worst case) and long hangs (in the best case). As ZSYSTEM and BREAK are neither isolated nor timed, YottaDB initiates TPNOTACID behavior for them immediately as it encounters them during execution in a final retry attempt (independent of ydb_tpnotacidtime). Rapidly repeating TPNOTACID messages are likely associated with live-lock, which means that a process is consuming critical resources repeatedly within a transaction, and is unable to commit because the transaction duration is too long to commit while maintaining ACID transaction properties.

ydb_tprestart_log_delta

ydb_tprestart_log_delta (gtm_tprestart_log_delta) specifies the frequency with which YottaDB reports transaction restarts to the syslog. A value of 1 means that every transaction restart is to be reported. If ydb_tprestart_log_delta is not defined, YottaDB initializes ydb_tprestart_log_delta to 0, meaning that no restarts are to be reported, regardless of the value of ydb_tprestart_log_first.

ydb_tprestart_log_first

ydb_tprestart_log_first (gtm_tprestart_log_first) specifies the initial number of transaction restarts which YottaDB should report before pacing subsequent transaction restart reports to the syslog using the ydb_tprestart_log_delta value. If ydb_tprestart_log_delta is defined and ydb_tprestart_log_first is not defined, YottaDB initializes ydb_tprestart_log_first to 0.

ydb_trace_gbl_name

ydb_trace_gbl_name (gtm_trace_gbl_name) enables YottaDB tracing at process startup. Setting ydb_trace_gbl_name to a valid global variable name instructs YottaDB to report the data in the specified global when a VIEW command disables the tracing, or implicitly at process termination. This setting behaves as if the process issued a VIEW "TRACE" command at process startup. However, ydb_trace_gbl_name has a capability not available with the VIEW command, such that if the environment variable is defined but evaluates to zero (0) or to the empty string, YottaDB collects the M-profiling data in memory and discards it when the process terminates (this feature is mainly used for in-house testing). Note that having this feature activated for processes that otherwise don't open a database file (such as GDE) can cause them to encounter an error.

ydb_treat_sigusr2_like_sigusr1

ydb_treat_sigusr2_like_sigusr1 when set to a non-zero numeric value, "yes" or "TRUE" (case-insensitive), or a leading substring of "yes" or "true", causes a YottaDB process to treat a USR2 signal just as it would a SIGUSR1 (by invoking $ZINTERRUPT mechanism). The default behavior is to ignore SIGUSR2.

ydb_treat_sigusr2_like_sigusr1 was added to YottaDB release r1.32.

ydb_trigger_etrap

ydb_trigger_etrap (gtm_trigger_etrap) provides the initial value for $ETRAP in trigger context; can be used to set trigger error traps for trigger operations in both yottadb and MUPIP processes.

ydb_xc_gpgagent

ydb_xc_gpgagent (GTMXC_gpgagent) specifies the location of gpgagent.tab. By default, YottaDB places gpgagent.tab in the $ydb_dist/plugin/ directory. ydb_xc_gpgagent is used by pinentry-gtm.sh and is meaningful only if you are using Gnu Privacy Guard version 2.

ydb_zdate_form

ydb_zdate_form (gtm_zdate_form) specifies the initial value for the $ZDATE ISV. The SET command can alter the value of $ZDATE in an active process.

ydb_zinterrupt

ydb_zinterrupt (gtm_zinterrupt) specifies the initial value of the $ZINTERRUPT intrinsic special variable which holds the code that YottaDB executes (as if it were the argument of an XECUTE command) when a process receives a signal from a MUPIP INTRPT command.

ydb_zlib_cmp_level

ydb_zlib_cmp_level (gtm_zlib_cmp_level) specifies the zlib compression level used in the replication stream by the source and receiver servers. By default, replication does not use compression.

ydb_zquit_anyway

ydb_zquit_anyway (gtm_zquit_anyway) specifies whether the code of the form QUIT <expr> execute as if it were SET <tmp>=<expr> QUIT:$QUIT tmp QUIT, where <tmp> is a temporary local variable in the YottaDB runtime system that is not visible to application code. This setting is a run-time setting, rather than a compiler-time setting. If ydb_zquit_anyway is defined and evaluates to 1 or any case-independent string or leading substring of "TRUE" or "YES", code of the form QUIT <expr> executes as if it were SET <tmp>=<expr> QUIT:$QUIT tmp QUIT. If ydb_zquit_anyway is not defined or evaluates to 0 or any case-independent string or leading substring of "FALSE" or "NO", YottaDB executes QUIT <expr> as specified by the standard.

ydb_zstep

ydb_zstep (gtm_zstep) specifies the initial value of $ZSTEP, which defines the ZSTEP action; if ydb_zstep is not defined, $ZSTEP defaults to "B".

ydb_ztrap_form/ydb_zyerror

ydb_ztrap_form (gtm_ztrap_form) and ydb_zyerror (gtm_zyerror) specify the behavior of error handling specified by $ZTRAP as described in the Error Processing chapter of the Programmer's Guide.

ydb_ztrap_new

ydb_ztrap_new (gtm_ztrap_new) specifies whether a SET $ZTRAP also implicitly performs a NEW $ZTRAP before the SET.

Some Things To Remember

There is a lot of information to digest. Here are some things to remember as you start with YottaDB and build your expertise.

For those of the following environment variables which are not set, ydb_env_set sets reasonable defaults: ydb_chset, ydb_dir, ydb_dist, ydb_etrap, ydb_gbldir, ydb_icu_version, ydb_log, ydb_procstuckexec, ydb_rel, ydb_repl_instance, ydb_retention, ydb_routines, ydb_tmp, and ydb_xc_* variables for installed plugins.

YottaDB recommends using the ydb_env_set script (or the ydb script which sources ydb_env_set) to set up an environment for YottaDB.

While creating an environment for multiple processes accessing the same version of YottaDB, bear in mind the following important points:

  • A YottaDB version has an associated gtmsecshr (located by $ydb_dist). If multiple processes are accessing the same YottaDB version, each process must use the same $ydb_tmp

  • YottaDB recommends setting $ydb_tmp to a temporary directory. The ydb_env_set script sets ydb_tmp to :code;`/tmp/yottadb/$ydb_rel` where $ydb_rel is the current YottaDB release, e.g., r1.36_x86_64.

Always set the same value of $ydb_tmp for all processes using the same YottaDB version. Having different $ydb_tmp for multiple processes accessing the same YottaDB version may prevent processes from being able to communicate with gtmsecshr and cause performance disruption.

Configuring and operating YottaDB with Unicode™ support (optional) for M

Data is stored in a database as byte sequences, and the database is agnostic about the interpretation of those byte sequences as characters. Mapping between bytes and characters, i.e., treating the bytes as single-byte characters or multi-byte characters, is done by application software. Whether to install YottaDB with Unicode support depends on applicate code.

An M process operates in either M mode (single-byte characters) or UTF-8 mode (multi-byte characters). Traditional M applications typically do not use Unicode, and UTF-8 mode need not be installed if YottaDB is to be used only for a traditional M application. As most other languages support Unicode strings by default, and as some YottaDB utility programs are written in M, YottaDD recommends installing YottaDB with Unicode support unless you know that the YottaDB installation is only to be used by traditional M code.

Running YottaDB

Refer to the M Programmers Guide to run YottaDB in M mode, and the Multi-Language Programmers Guide to run YottaDB programs.

Configuring hugepages for YottaDB on Linux

Hugepages are a Linux feature that may improve the performance of YottaDB applications in production. Hugepages create a single page table entry for a large block (typically 2MiB) of memory in place of hundreds of entries for many smaller (typically 4KiB) blocks. This reduction of memory used for page tables frees up memory for other uses, such as file system caches, and increases the probability of TLB (translation lookaside buffer) matches - both of which can improve performance. The performance improvement related to reducing the page table size becomes evident when many processes share memory as they do for global buffers, journal buffers, and replication journal pools. Configuring hugepages on Linux may help improve:

  • Shared memory performance: When your YottaDB database uses journaling, replication, and the BG access method.

  • Process memory performance: For your process working space and dynamically linked code.

    Note

    At this time, hugepages have no effect for MM databases; the text, data, or bss segments for each process; or for process stack.

While YottaDB recommends you configure hugepages for shared memory, you need to evaluate whether or not configuring hugepages for process-private memory is appropriate for your application. Having insufficient hugepages available during certain commands (for example, a JOB command - see complete list below) can result in a process terminating with a SIGBUS error. This is a current limitation of Linux. Before you use hugepages for process-private memory on production systems, YottaDB recommends that you perform appropriate peak load tests on your application and ensure that you have an adequate number of hugepages configured for your peak workloads or that your application is configured to perform robustly when processes terminate with SIGBUS errors.

The following YottaDB features fork processes and may generate SIGBUS errors when hugepages are not available - JOB, OPEN a PIPE device, ZSYSTEM, interprocess signaling that requires the services of gtmsecshr when gtmsecshr is not already running, SPAWN commands in DSE, GDE, and LKE, argumentless MUPIP RUNDOWN, and replication-related MUPIP commands that start server processes and/or helper processes. Should increasing the available hugepages require a reboot, an interim workaround is to unset the environment variable HUGETLB_MORECORE for YottaDB processes until you are able to reboot or otherwise make available an adequate supply of hugepages.

Consider the following example of a memory map report of a Source Server process running at peak load:

$ pmap -d 18839
18839: /usr/lib/yottadb/r120/mupip replicate -source -start -buffsize=1048576 -secondary=melbourne:1235 -log=/var/log/.yottadb/mal2mel.log -instsecondary=melbourne
Address   Kbytes Mode Offset   Device Mapping
--- lines removed for brevity -----
mapped: 61604K writeable/private: 3592K shared: 33532K
$

Process id 18839 uses a large amount of shared memory (33535K) and can benefit from configuring hugepages for shared memory. Configuring hugepages for shared memory does not cause a SIGBUS error when a process does a fork. For information on configuring hugepages for shared memory, refer to the "Using hugepages" and "Using hugepages for shared memory" topics below. SIGBUS errors only occur when you configure hugepages for process-private memory; these errors indicate you have not configured your system with an adequate number of hugepages. To prevent SIGBUS errors, you should perform peak load tests on your application to determine the number of required hugepages. For information on configuring hugepages for process-private memory, refer to the "Using hugepages" and "Using hugepages for process working space" sections.

As application response time can be adversely affected if processes and database shared memory segments are paged out, YottaDB recommends configuring systems for use in production with sufficient RAM so as to not require swap space or a swap file. While you must configure an adequate number of hugepages for your application needs as empirically determined by benchmarking/testing and there is little downside to a generous configuration to ensure a buffer of hugepages available for workload spikes, an excessive allocation of hugepages may affect system throughput by reserving memory for hugepages that could otherwise be used by applications that cannot use hugepages.

Using Hugepages

Prerequisites

Notes

A CPU running a Linux kernel with hugepages enabled.

All currently Supported Linux distributions appear to support hugepages; to confirm, use the command: grep hugetlbfs /proc/filesystems which should report: nodev hugetlbfs. If it does not, refer to your Linux distribution's documentation to make hugepages available for use.

A sufficient number of hugepages available.

To reserve hugepages boot Linux with the hugepages=<num_pages> kernel boot parameter; or, shortly after bootup when unfragmented memory is still available, with the command: hugeadm --pool-pages-min DEFAULT:<num_pages>. For subsequent on-demand allocation of hugepages, use: hugeadm --pool-pages-max DEFAULT:<num_pages> These delayed (from boot) actions do not guarantee availability of the requested number of hugepages; however, they are safe as, if a sufficient number of hugepages is not available, Linux simply uses traditional sized pages.

Using Hugepages for Shared Memory

To use hugepages for shared memory (journal buffers, replication journal pool, global buffers and M code executing from shared memory):

  • Permit YottaDB processes to use hugepages for shared memory segments (YottaDB recommends option 1 below, since the ext4 and xfs filesystems both support extended attributes) using one of the following, both of which require administrative privileges:

    1. Set the CAP_IPC_LOCK capability for your yottadb, dse, and mupip processes, with a command such as:

      setcap 'cap_ipc_lock+ep' $ydb_dist/yottadb

      Non-M applications that access YottaDB databases, through a non-M language API or call-ins should also set the capability on the top level executable files.

    1. Permit the group used by YottaDB processes that use huge pages with the following command

      echo <gid> >/proc/sys/vm/hugetlb_shm_group

      Note

      The /proc/sys/vm/hugetlb_shm_group setting needs to be preserved on reboot, e.g., in /etc/sysctl.conf or a startup script.

  • Set the environment variable ydb_hugetlb_shm for each process to yes. The older technique, setting the environment variable HUGETLB_SHM set to yes, potentially with LD_PRELOAD set to the path to libhugetlbfs.so also works, but YottaDB will not know that hugepages are being used for shared memory in that case.

Note

Since the memory allocated by Linux for shared memory segments mapped with hugepages is rounded up to the next multiple of hugepages, there is potentially unused memory in each such shared memory segment. You can therefore increase any or all of the number of global buffers, journal buffers, and lock space to make use of this otherwise unused space. You can make this determination by looking at the size of shared memory segments using ipcs. Contact YottaDB support for a sample program to help you automate the estimate. Transparent hugepages may further improve virtual memory page table efficiency. Some supported releases automatically set transparent_hugepages to "always"; others may require it to be set at or shortly after boot-up. Consult your Linux distribution's documentation.

Using hugepages for YottaDB process working space

To use hugepages for process working space and M code running in process-private memory:

  • Set the environment variable HUGETLB_MORECORE for each process to yes.

If you enable hugepages for all applications (by setting HUGETLB_MORECORE and HUGETLB_SHM as discussed above in /etc/profile and/or /etc/csh.login), you may find it convenient to suppress warning messages from common applications that are not configured to take advantage of hugepages by also setting the environment variable HUGETLB_VERBOSE to zero (0).

Refer to the documentation of your Linux distribution for details. Other sources of information are:

Configuring the Restriction Facility

Post installation, a system administrator can optionally add a restrict.txt file in $ydb_dist to restrict the use of certain YottaDB facilities to a group-name. The owner and group for $ydb_dist/restrict.txt can be different from those used to install YottaDB.

The file may contain zero or more of the following case-insensitive lines, in the specified format, in any order:

A{D|L|M|PD|ZA}_ENABLE:[comma-separated-list-of-options]:{path-to-sock-file|host:port}[:tls-id]
BREAK[:<group-name>]
CENABLE[:<group-name>]
DIRECT_MODE[:<group-name>]
DSE[:<group-name>]
HALT[:<group-name>]
LIBRARY:[<group-name>]
LKE:[<group-name>]
LKECLEAR:[<group-name>]
LOGDENIALS[:<group-name>]
PIPE_OPEN[:<group-name>]
TRIGGER_MOD[:<group-name>]
ZBREAK[:<group-name>]
ZCMDLINE[:<group-name>]
ZEDIT[:<group-name>]
ZHALT[:<group-nam>]
ZLINK[:<group-nam>]
ZROUTINES[:<group-nam>]
ZRUPDATE[:<group-nam>]
ZSYSTEM[:<group-name>]

If the file $ydb_dist/restrict.txt does not exist, YottaDB does not restrict any facilities.

If the file exists, a process that has:

  • write authorization to restrict.txt has no restrictions;

  • no access to restrict.txt is restricted from all facilities for which YottaDB supports a restriction (the above list); and

  • read-only access to restrict.txt is restricted from any listed facility unless it is a member of the group specified in the optional group-id following the facility name.

In addition, a process that has read-only access to restrict.txt is restricted from all facilities if any of the below conditions are met.

  • Any non-empty lines that do not match the above format.

  • Any <group-name> mentioned after the facility name is not a valid group name in the system/host.

Restrictions apply as follows:

YottaDB Facility

Behavior

APD_ENABLE

YottaDB supports the ability to log actions initiated from a principal device including M commands typed interactively, or piped in by a script or redirect, from the principal device ($PRINCIPAL) and/or any information entered in response to a READ from $PRINCIPAL. An action initiated from $PRINCIPAL executes as usual when Audit Principal Device is disabled, which it is by default. However, when Audit Principal Device is enabled, YottaDB attempts to send the action out for logging before acting on it. Additionally, the $ZAUDIT Intrinsic Special Variable (ISV) provides a Boolean value that indicates whether Audit Principal Device is enabled. See the Audit Principal Device section below for details.

AD_ENABLE

Enables the logging of DSE commands from the shell and DSE prompt.

AL_ENABLE

Enables the logging of LKE commands from the shell and LKE prompt.

AM_ENABLE

Enables the logging of MUPIP commands from the shell and MUPIP prompt.

AZA_ENABLE

Enables the use of the $ZAUDITLOG() function. When LGDE is specified as the keyword for AZA_ENABLE, GDE logs GDE commands.

BREAK

YottaDB ignores any BREAK command.

CENABLE

The process acts as if $ydb_nocenable is TRUE and ignores any CENABLE deviceparameter.

DIRECT_MODE

yottadb -direct terminates immediately with a RESTRICTEDOP error.

DSE

DSE terminates immediately with a RESTRICTEDOP error.

HALT

HALT results in a RESTRICTEDOP error.

LIBRARY

Any attempt to load an external library produces a RESTRICTEDOP error.

Libary restrictions apply to third party libraries loaded by YottaDB directly, and those loaded at the behest of YottaDB applications. This restriction allows administrators to control which libraries YottaDB is able to load at run-time. To allow library loading, place symbolic links to the desired libraries in $ydb_dist/plugin.

LKE

Any invocation of the LKE utility produces a RESTRICTEDOP error.

LKECLEAR

Any invocation of the LKE CLEAR command results in a RESTRICTEDOP error.

LOGDENIALS

Limit logging to processes belonging to a specific group.

YottaDB normally logs a number of errors related to permissions and access using the syslog facility. The YottaDB restriction LOGDENIALS provides a facility for disabling this logging on a group basis. If this mechanism is not used, the logging occurs for all YottaDB processes If the restriction is used, logging occurs for the specified group only. YottaDB supports group names using the POSIX Portable Filename Character Set which includes characters from [A-Z], [a-z], [0-9], ., _ , and -.

PIPE_OPEN

OPEN of a PIPE device produces a RESTRICTEDOP error.

TRIGGER_MOD

A $ZTRIGGER() or MUPIP TRIGGER that attempts a change or delete produces a RESTRICTEDOP error; in addition, while executing code within a trigger, ZBREAK results in a RESTRICTEDOP error, and both ZBREAK and ZSTEP actions are ignored.

ZBREAK

Any ZBREAK produces a RESTRICTEDOP error.

ZCMDLINE

YottaDB returns an empty string for all references to $ZCMDLINE.

ZEDIT

ZEDIT produces a RESTRICTEDOP error.

ZHALT

ZHALT produces a RESTRICTEDOP error.

ZLINK

An explicit ZLINK produces a RESTRICTEDOP error.

ZROUTINES

A SET of $ZROUTINES produces a RESTRICTEDOP error.

ZRUPDATE

ZRUPDATE produces a RESTRICTEDOP error.

ZSYSTEM

ZSYSTEM produces a RESTRICTEDOP error.

Note that restricting $ZCMDLINE prevents commands like: yottadb -run %XCMD 'for read x xecute x' which can act as substitutes for Direct Mode.

In order to limit pathological looping from restricted HALT or ZHALT, if a YottaDB process issues a second occurrence of the restricted command within half a second, the process terminates after sending a fatal error to both the principal device and the syslog, and also produces a YDB_FATAL* context file, but no core file. With these restrictions in place, a process should terminate with, for example: ZGOTO 0. Note that with or without a restriction, executing these commands as part triggered logic on a replicating instance may cause the Update Server to terminate and thereby stop replication.

ZSYSTEM and PIPE OPEN command restriction facility

The YottaDB restriction mechanism recognizes the following lines:

ZSYSTEM_FILTER[:M labelref]
PIPE_FILTER[:M labelref]

The labelref must include a routine name. If a process is restricted by a ZSYSTEM or PIPE_OPEN line in the restrictions file, that restriction takes precedence over the corresponding filter restriction. Otherwise when a process is subject to these restrictions, YottaDB inserts an invocation of the labelref prior to the restricted command, passing a string containing the argument to the ZSYSTEM command or the command deviceparameter of the PIPE OPEN. The path to the filter routine must be included in $zroutines. YottaDB recommends that the filter routine be placed in a location with restricted access such as $ydb_dist. If the filter invocation return is -1, YottaDB produces a RESTRICTEDOP error, otherwise, it executes the command using the returned string via output parameters as a possibly identical replacement for the original string. Note that because ZSYSTEM and OPEN are not Isolated actions, YottaDB recommends against their use within a TP transaction. Filters will also increment the nested level of call-ins. A recursive filter invocation produces a NOFILTERNEST error. YottaDB reports all filter errors to the syslog accompanied by a COMMFILTERERR.

An example restrict file for this:

cat $ydb_dist/restrict.txt
ZSYSTEM_FILTER:^filterzsy
PIPE_FILTER:^filterzsy

The actual filter routine:

  filterzsy(inarg,outarg);
  if ""=inarg set outarg="-1;must provide a command" quit
  for i=1:1 set arg=$piece(inarg,";",i) quit:""=arg  do  quit:$data(outarg)
  . for  quit:$zchar(9,32)'[$extract(arg)  set arg=$extract(arg,2,9999)
  . set cmd=$piece(arg," ")
  . for restrict="sudo","cd" if cmd=restrict set outarg="-1;command "_restrict_" not permitted" quit
  . quit:$data(outarg)
  . if "echo"=cmd set $piece(arg," ")="echo #",$piece(inarg,";",i)=arg    ;example of modification
  set:'$data(outarg) outarg=inarg
  quit +outarg

Filter execution starts with $STACK=1 ($ZLEVEL=2).

The commands, Intrinsic Special Variables, and functions whose behavior changes in the context of a filter invocation are below:

  • ZGOTO 0 (zero) returns to the processing of the restricted command as does ZGOTO 1 (one) with no entryref, while ZGOTO 1:entryref replaces the originally invoked filter and continues filter execution.

  • $ZTRAP/$ETRAP NEW'd at level 1.

  • $ZLEVEL initializes to one (1) in GTM$CI, and increments for every new stack level.

  • $STACK initializes to zero (0) in GTM$CI frame, and increments for every new stack level.

  • $ESTACK NEW'd at level one (1) in GTM$CI frame.

  • $ECODE/$STACK() initialized to the empty string at level one (1) in GTM$CI frame.

After the filter completes, YottaDB restores the above to their values at the invocation of the filter.

Audit Logging Facility

To enable the audit logging facility, add a line in the following format to $ydb_dist/restrict.txt:

A{D|L|M|PD|ZA}_ENABLE:[comma-separated-list-of-options]:{path-to-sock-file|host:port}[:tls-id]

APD_ENABLE enables the logging of all code entered from Direct Mode and optionally any input entered on the principal device ($PRINCIPAL). AD_ENABLE, AL_ENABLE and AM_ENABLE enable the logging of all DSE, LKE and MUPIP commands entered from the shell and utility prompt. AZA_ENABLE enables the logging for arguments to the $ZAUDITLOG() function.

  • The optional "comma-separated-list-of-options" can consist of zero or more of these options:

    • LGDE - Enable logging of all GDE commands from shell and GDE prompt. The LGDE option only applies to AZA_ENABLE.

    • RD - Enable logging of all responses READ from $PRINCIPAL in addition to that entered at the Direct Mode prompt. This option is more comprehensive and captures input that might be XECUTEd, but depending on your application architecture may significantly increase the amount of logged information. The RD option only applies to APD_ENABLE.

    • TLS - Enables TLS connectivity between YottaDB and the logger; this option requires a tls-id and host information (e.g. IP/port or hostname/port).

  • The "path-to-sock-file" is the absolute path of the UNIX domain socket file for connecting to the logger.

  • The "host" is the hostname or numeric IPv4/IPv6 address of the logger; numeric IP addresses must be enclosed in square brackets (i.e. [ and ]).

  • The "port" is the port number the logger listens on.

  • You can specify the same or different {path-to-sock-file|host:port} for any audit logging facility.

  • The optional "tls-id" is the label of the section within the YottaDB configuration file that contains TLS options and/or certificates for YottaDB to use; logging ignores any "tls-id" if the "TLS" option is not specified.

If parsing the "A*_ENABLE" line in the restriction file or initializing logger information fails, YottaDB enforces all restrictions (default restriction file behavior).

Examples:

APD_ENABLE::/path/to/sock/file/audit.sock

Adding this line to the restriction file enables APD. YottaDB connects with the logger via UNIX domain socket using the domain socket file "/path/to/sock/file/audit.sock" and sends all Direct Mode activity from $PRINCIPAL to logger.

APD_ENABLE:RD:[123.456.789.100]:12345

Adding this line to the restriction file enables APD. YottaDB connects with the logger (listening on port 12345 at the IPv4 address 123.456.789.100) via a TCP socket and sends all Direct Mode and READ activities from $PRINCIPAL to logger.

APD_ENABLE::loggerhost:56789

Adding this line to the restriction file enables APD. YottaDB connects with the logger (listening on port 56789 at the hostname "loggerhost") using a TCP socket and sends all Direct Mode activities from $PRINCIPAL to logger.

APD_ENABLE:TLS,RD:[1234:5678:910a:bcde::f:]:12345:clicert

Adding this line to the restriction file enables APD. YottaDB connects with the logger (listening on port 12345 at the IPv6 address 1234:5678:910a:bcde::f:) via TLS socket. YottaDB configures its TLS options for APD based on the contents within the section of the configuration file labeled "clicert". YottaDB sends all Direct Mode and READ activities from $PRINCIPAL to logger.

AD_ENABLE::/path/to/sock/file/audit.sock

Adding this line to the restriction file enables the logging of all DSE commands. YottaDB connects with the logger via a UNIX domain socket using the domain socket file /path/to/sock/file/audit.sock and sends all DSE activity to the logger.

AL_ENABLE::/path/to/sock/file/audit.sock

Adding this line to the restriction file enables the logging of all LKE commands. YottaDB connects with the logger via a UNIX domain socket using the domain socket file /path/to/sock/file/audit.sock and sends all LKE activity to the logger.

AM_ENABLE::/path/to/sock/file/audit.sock

Adding this line to the restriction file enables the logging of all MUPIP commands. YottaDB connects with the logger via a UNIX domain socket using the domain socket file /path/to/sock/file/audit.sock and sends all MUPIP activity to the logger.

Logging

A "logger" is a separate server-like program responsible for receiving and logging information from YottaDB audit logging facilities (A*_ENABLE). A logger program can run in the foreground or background. You can use the same logger program or specify different logger programs for each audit logging facilities.

YottaDB includes sample logger programs that you can download and adapt to your needs. dm_audit_listener.zip contains three sample logger programs that can be used for logging. The difference among the three is the socket connection type used to communicate with YottaDB. The dm_audit_unix_listener.c, dm_audit_tcp_listener.c, and dm_audit_tls_listener.c use UNIX domain, TCP, and TLS sockets respectively to communicate with YottaDB. Choose the logger program based on the connection type specified in the A*_ENABLE entry in $ydb_dist/restrict.txt. The zip file contains a Makefile to help with compilation.

You can unzip dm_audit_listener.zip and run make to compile all three logger programs. To compile an individual program, just run make with the name of the program, e.g., make dm_audit_tcp_listener. To compile dm_audit_tls_listener requires the OpenSSL package.

Running the logger programs:

  • Run the UNIX domain socket logger as follows:

    dm_audit_unix <logfile> <sockfile>

    • <logfile> is the path to the output audit log.

    • <sockfile> is the path to the Unix domain socket file.

  • Run the TCP logger as follows:

    dm_audit_tcp <logfile> <portno>

    • <logfile> is the path to the output audit log.

    • <portno> is the port number on which to listen.

  • Run the TLS logger as follows:

    dm_audit_tls_listener <logfile> <portno> <certfile> <privkeyfile> <passphrase> [-clicert] [-cafile <CAfilepath>] [-capath <CApath>]

    • <logfile> is the path to the output audit log.

    • <portno> is the port number on which to listen.

    • <certfile> is the path of the TLS certificate to use.

    • <privkeyfile> is the parth to the file with the private key.

    • <passphrase> is the password pr passphrase for the certificate / key.

    • If the -clicert option is present, the logger asks the YottaDB process for a client certificate.

    • If the -cafile option is specifed, <CAfilepath> is the path to a file of CA certificates in PEM format for verification purposes.

    • If the capath opton is specified, <CApath> is the path a directory of CA certificate files in PEM format for verification purposes.

Logger Message Format:

The seven fields in the message areseparated by semicolons (';'), and contain information on the to-be-logged activity. Each to-be-logged message sent to the logger from YottaDB has the following format:

dist=<path>; src={0|1|2|3|4|5|6}; uid=<uid>; euid=<euid>; pid=<pid>; tty=<ttyname> command=<text>

  • The dist field, shows the path to location of the sender/user's $ydb_dist (or top level program, if not an M program).

  • The src field shows zero (0) for input from unknown source, one (1) for Direct Mode input; two (2) for READ input from $PRINCIPAL or when the standard input to the Direct Mode prompt is not froma terminal (commands entered via HEREDOCs from a shell script, or a pipe); three (3) for MUPIP commands; four(4) for $ZAUDITLOG() and GDE; five (5) for LKE; and six (6) for DSE.

  • The next three fields (uid, euid, and pid) show respectively the user ID, effective user ID, and process ID of the process that sent the message, all in decimal.

  • The tty field shows the stdin for the process. If the stdin at process startup is not a terminal device, the field is tty=0.

  • The command field is the input provided on the YottaDB side.

Examples:

dist=/usr/local/lib/yottadb/r202; src=1; uid=112233445; euid=112233445; pid=987654; tty=/dev/pts/0; command=write "Hello world",!
dist=/usr/local/lib/yottadb/r202; src=2; uid=998877665; euid=998877665; pid=123456; tty=/dev/pts/1; command=read num
dist=/usr/local/lib/yottadb/r202; src=2; uid=998877665; euid=998877665; pid=123456; tty=/dev/pts/1; command=7

This example demonstrates the audit logging with APD_ENABLE:RD facility. Logging activity shows that PID 987654 and PID 123456 ran two Direct Mode commands - write "Hello world",! and read num. The response from PID 123456 for read num was 7.