03-17-2016 02:37 AM
Hi to All,
I wanted to ask You a question.
In our backup infrastructure, we Installed on a Windows server 2008r2 the OPS center 188.8.131.52 (NBU 184.108.40.206 on Solaris 10).
The ops db grew to approximately 100GB:
symcOpscache.db 59756536 KB
symcopsscratchdb.db 168 KB
symcsearchdb.db 1192 KB
vxpmdb.db 43034392 KB
I wanted to know if there is a way to shrink the DB without compromising the usability of the product.
Sorry for my english
03-17-2016 11:52 AM
I don't think you can shrink it, its size would be determined by the amount of data its collecting, and the retention you're keeping in opscenter (1 year, 2 years, etc).
03-21-2016 09:10 PM
Have you ever done a dbdefrag.bat ?
Note that your symcOpscache is the largest among those files, but the main DB file is vxpmdb. It is quite usual that symcOpscache grows bigger than the main one.
Anyway, I would suggest this in sequence:
1) Confirm you have enabled "data purge" and the days of purge is not set too high (In GUI: Settings -> NetBackup -> Data Purge)
2) Perform a dbdefrag.bat (this will also backup a copy of DB before the defrag) : http://www.veritas.com/docs/000068155
3) If after this, you did not see any improvement, consider upgrade to latest version such as 220.127.116.11 or 7.7.1
4) Open a support case if it's still not good.
03-24-2016 07:04 AM
That symcOpscache.db size doesnt look normal, if i recall correctly this should only have the latest 30days of OpsCenter data. I'd recommend opening a support case to find out whats causing the bloat.
03-24-2016 08:13 AM
Hi to All and thanks for answers,
Unfortunately, the "data purge" is deactivated for company policies.
With the help of the support we are providing to implement the dbdefrag.
The first attempt has failed due to lack of space, despite the db is on a disk with 101GB of available space.
I still do not understand why, considering that it is recommended to have free space equal to twice the size of the db.
for the symcOpscache.db files size, the support did not find anything anomalous...
03-29-2016 02:00 PM
You should really challenge that company policy. Unfortunately, OpsCenter's DB software just doesnt work very well with such large databases. As such,Veritas really doesnt recommend disabling purge forever. So even if you get this problem fixed you will eventually run into trouble, in the meantime you have a lot of pain to look forward to - upgrades, defrags, migrations, etc. will all be much more difficult with a DB of this size.
You can instead take regular backups of the DB, keep those as long as needed, and if old data is needed do a restore. I like to keep a powered down OpsCenter VM with the same version binaries as production for cases like these.
Also - free space equal to twice the size of the DB would mean that you need 200GB Free (told you its gonna be painful :))