Red Hat Training

A Red Hat training course is available for Red Hat Directory Server

20.5. Monitoring the Local Disk for Graceful Shutdown

When the disk space available on a system becomes too small, the Directory Server process (slapd) crashes. Any abrupt shutdown runs the risk of corrupting the database or losing directory data.
It is possible to monitor the disk space available to the slapd process. A disk monitoring thread is enabled using the nsslapd-disk-monitoring configuration attribute. This creates a monitoring thread that wakes every ten (10) seconds to check for available disk space in certain areas.
If the disk space approaches a defined threshold, then the slapd begins a series of steps (by default) to reduce the amount of disk space it is consuming:
  • Verbose logging is disabled.
  • Access logging and error logging are disabled.
  • Rotated (archived) logs are deleted.

Note

Error log messages are always recorded, even when other changes are made to the logging configuration.
If the available disk space continues to drop to half of the configured threshold, then the slapd begins a graceful shut down process (within a grace period); and if the available disk space ever drops to 4KB, then the slapd process shuts down immediately. If the disk space is freed up, then the shutdown process is aborted, and all of the previously disabled log settings are re-enabled.
By default, the monitoring thread checks the configuration, transaction log, and database directories. An additional attribute (nsslapd-disk-monitoring-logging-critical) can be set to include the logs directory when evaluating disk space.
Disk monitoring is disabled by default, but it can be enabled and configured by adding the appropriate configuration attributes to the cn=config entry. Table 20.1, “Disk Monitoring Configuration Attributes” lists all of the configuration options.
  1. Using ldapmodify, add the disk monitoring attributes. At a minimum, turn on the nsslapd-disk-monitoring attribute to enable disk monitoring. The default threshold is 2MB; this can be configured (optionally) in the nsslapd-disk-monitoring-threshold attribute.
    For example: 
    # ldapmodify -D "cn=Directory Manager" -W -x
dn: cn=config
changetype: modify
add: nsslapd-disk-monitoring
nsslapd-disk-monitoring: on
-
add: nsslapd-disk-monitoring-threshold
nsslapd-disk-monitoring-threshold: 3000000
-
add: nsslapd-disk-monitoring-grace-period
nsslapd-disk-monitoring-grace-period: 20
  2. Restart the Directory Server to load the new configuration. 
    [root@server ~]# systemctl restart dirsrv.target

Table 20.1. Disk Monitoring Configuration Attributes

Configuration Attribute Description
nsslapd-disk-monitoring Enabled disk monitoring. This is the only required attribute, since the other configuration options have usable defaults.
nsslapd-disk-monitoring-grace-period Sets a grace period to wait before shutting down the server after it hits half of the disk space limit. This gives an administrator time to address the situation. The default value is 60 (minutes).
nsslapd-disk-monitoring-logging-critical Sets whether to shut down the server if the log directories pass the halfway point set in the disk space limit. This prevents the monitoring thread from disabling audit or access logging or from deleting rotated logfiles.
nsslapd-disk-monitoring-threshold Sets the amount of disk space, in bytes, to use to evaluate whether the server has enough available disk space. Once the space reaches half of this threshold, then the server begins a shut down process. The default value is 2000000 (2MB).