2013-06-20 01:13 PM - edited 2015-12-18 12:31 AM
We can configure the SC server to wait for an agent quiescing large Domino databases, but is there a way to configure the agent so that it will wait or retry access to a database locked for
some maintenance task for instance?
Thanks for comments and suggestions.
2013-06-20 05:31 PM
The Domino plug-in uses the Domino APIs to put databases into backup mode.
This all happens pretty quickly, so while it would be possible to retry locked databases it likely wouldn't have much of an effect.
Database locks tend to last longer than the few minute the backup start operation takes.
We would not want to hold all operations until locks are released as this could take an unknown about of time and could lead to larger problems.
I can add an enhancement request to see if we can figure out some options here, but off the top of my head there isn't much we can do in the plug-in at the moment.
You could optionally add a PRE_APP_QUIESCE command to run Domino maintenance tasks before the backup start begins.
Consult your Domino team (or perhaps you are your Domino team ) for the correct commands.
This still wouldn't guarantee that there wouldn't be any locks - You may need to figure out what is causing the locks and resolve the underlying issue.
2013-06-23 01:40 PM
Thanks a lot John,
I will suggest indeed to synchronize the compaction task using the PRE_APP_QUIESCE directive inside the snapcreator configuration file.
Thanks for the suggestion
2013-06-23 01:54 PM
Two additional notes:
1) if you decide to run compact as a PRE_APP_QUIESCE task, be sure to increase the agent timeout by a significant value to allow enough time for all compact operations to complete. I've not run compact as a PRE_APP_QUIESCE, so I am not sure of the exact behavior.
Domino will either:
a) simply accept the command and return a zero error code immediately - this means that the compact will start but Snap Creator will continue on normally while the compact is in process. You may have a database or two locked as a result.
b) Domino will wait for the compact operation to complete and then return the zero error code. This may mean that the Snap Creator Agent is alive for a long period of time. This will increase the overall time Snap Creator runs and will necessitate a long timeout period on the agent.
2) Please pay attention to the type of compact operation you are running in Domino. Generally speaking (knowing absolutely nothing about the environment) we recommend compact -B -S XX with XX being either 10 or 15. This means that compact will only run when there is greater than XX% (10 or 15 suggested) whitespace in a database.
If you run frequent compact -c operations you will generate very large snapshot sizes. There is nothing wrong with an occasional compact -c, but you should not perform this operation very frequently. This is detailed in NetApp TR-3917: http://www.netapp.com/us/media/tr-3917.pdf
Please feel free to let me know if you have additional questions.