DOCUMENTATION: What are some of the situations where is it necessary to restart daemons / services in VERITAS NetBackup (tm) 6.0?
http://support.veritas.com/docs/281317
Details:
Manual: VERITAS NetBackup (tm) 6.0 Troubleshooting Guide for UNIX and Windows
Modification Type: Supplement
Modification:
The following general guidelines can be used to determine if a restart of NetBackup is required after making changes to the configuration on a NetBackup server or client.
The following terminology is used in this document:
Daemon: On UNIX, a daemon is a process that keeps running in the background
Service: On Windows, a service is specifically registered with Windows and interacts with a service management framework
ltid: the "Media Manager device daemon"daemon on UNIX. On Windows, this is know as the NetBackup Device Manager service
Legacy logging: debug logs traditionally used in NetBackup, enabled by creating a directory/folder
Unified logging: debug logs introduced in NetBackup 6.0 under the VxUL (Unified Logging) framework
EMM: Enterprise Media Manager; centralized daemon/service that handles resource management in NetBackup 6.0
The following are some common situations where it may be necessary to restart some or all of the NetBackup 6.0 daemons:
I. Changing the storage unit configuration
When adding, removing, or modifying a storage unit in NetBackup 6.0, no restarts are necessary. Changes will take effect for the next backup or restore that uses the storage unit. Any existing backups will not be updated with the changes if they were started before the changes were made.
Adding: No daemons need to be restarted
Modifying: No daemons need to be restarted
Deleting: No daemons need to be restarted
II. Changing the device (robot/drive) configuration
When adding, removing or modifying a device, some restarts may be necessary depending on the situation.
Adding: The ltid daemon on all the media servers that access the device need to be restarted.
Modifying: Daemon restart (on one or more than one media server) may or may not be required depending on the device attributes that are changed.
- If an attribute like drive cleaning frequency is changed, then no daemons need to be restarted.
- If an attribute like drive type or robot number is changed, then ltid on all the media servers that access the device need to be restarted.
- If the status column of vmoprcmd displays RESTART for a device path, then the media server corresponding to the device path needs to be restarted. In addition, if the robot attributes are changed, then ltid on the server acting as the robot control host may have to be restarted.
Deleting: No daemons need to be restarted.
III. Adding/modifying/deleting a media server
Adding: After NetBackup has been installed, the daemons on that media server need to be started if not already done so by the installation.
Modifying: No daemons need to be restarted.
Deleting: The media server should be deleted from the EMM server using the nbemmcmd -deletehost command. No daemons need to be restarted on other media servers. Additional steps are required to properly decommission a media server. Contact Symantec Enterprise Technical Services for information on how to properly decommission a media server.
IV. Using the FORCE_RESTORE_MEDIA_SERVER option in the bp.conf file
To override the media server to use for a restore, the FORCE_RESTORE_MEDIA_SERVER option can be added to the bp.conf file. This does not require any daemons be restarted. This option will take effect for any future restores. Any restores started prior to making the change will not be affected.
Adding: No daemons need to be restarted.
Modifying: No daemons need to be restarted.
Deleting: No daemons need to be restarted.
V. Changing policy and schedule configuration
The nbpem daemon in NetBackup 6.0 will handle any updates to policy information. The nbpem daemon will receive and process updates based on the Policy Update interval set in the Global Attributes of the master server. The default value is to check for and process updates to polices every ten minutes.
Adding: No daemons need to be restarted.
Modifying: No daemons need to be restarted.
Deleting: No daemons need to be restarted.
VI. Adding or removing a client
Adding: After NetBackup has been installed, the daemons on the client need to be started if not done so by the installation.
Deleting: The client being deleted should be removed from any active policies. No daemons need to be restarted.
VII. Enabling NetBackup logging (legacy or vxlogs) or changing the verbosity
The daemons for which logging is being turned on/off need to be restarted.
Unified logging (VxUL): These logs will be created automatically in /usr/openv/logs. The vxlogcfg command can be used to change logging levels for unified logging. Changes to the unified log settings, such as the DebugLevel, will require restarting NetBackup daemons.
Legacy logging:
- Persistent daemons such as bprd, bpjobd, bpdbm, bpcompatd, nbproxy, etc. will need to be restarted when creating or changing the VERBOSE logging level in the bp.conf file.
- Other NetBackup processes that are called only when needed such as bpbrm, bptm, bpcd, bpbkar, etc. will not need to be restarted when creating or changing the logging level. The changes will take effect the next time the process is started. Note: This will not affect any processes that were started before the change was made to the logging level.
VIII. Patch installation:
After the NetBackup patch has been installed, the daemons on the host on which the patch was installed need to be started if not already done so by the installation. For a silent installation, it is necessary to look at the /us/openv/tmp/install_trace log to determine whether a reboot is needed.
IX. Reboot on upgrade
Reboot only if prompted to do so by the installation
For a silent installation, look at the /us/openv/tmp/install_trace log to determine whether a reboot is neededMessage was edited by:
Bob Stump
