I only have 2 hosts and 30 Virtual machines and my DBA contacted me yesterday that my database grew 5 GB last week, and after only 2 months is up to 20GB already. I dont think I have changed logging from the default settings. Can anyone point me at a document which describes all the default logging and performance gathering settings for virtual center? I guess it would also help if there is anything describing what I can do to reduce the size also.
Here's the two large tables in the database:
dbo.VPX_EVENT_ARG | 24,919,168 | 9,552,608 | 5,200,368 | 4,350,680 | 1,560 | ||
dbo.VPX_EVENT | 16,308,312 | 8,173,432 | 4,509,256 | 3,662,912 | 1,264 |
Thanks
Mike
Hi,
It appears that your environment is pumping up with events - This is not normal behaviour. I have had this issue when I installed a tool called VCplus. A very handy tool that adds columns to your VI client environment to show disk usage per VM, and wheter a snapshot is in place. However, every time it updated those columns, an event was written too. My database had grown to a staggering 12GB (and only two hosts)....!!
So I would definitely check out the events in VirtualCenter - changes are you see events being logged every minute or so filling up your Database. Just click (from the VI client) on "hosts and clusters", task and events tab, and then the "events" button. That should give you a clue on what is happening.
I have been fooling around with SQL in order to remove all events older than 30 days, but in the end it did not reduce the database far enough for some reason, and I decided to start over (after removing VCplus that is )
That makes sense, we were tryng to configure SNMP alerts for events in virtual center to go to openview, and set the larm thresholds very low, and left them there a week, but that does not seem long enough to explain my whole issue, how long do events stay in virtual center? I dont see many there, they must stay longer in the DB.
Logging level can also increase the DB size. Check if it is set to high and reduce it if not needed.
Also read the following link, has useful info http://blog.scottlowe.org/2008/09/18/po2061-vmware-virtualcenter-25-database-best-practices/
is it normal to see root logging in and out multiple times every second? I think that may be what is killing my database size, I could not export it, and did not know where to look in virtual center server to find log file, I am attaching a screen shot of the activity. I see it at the host level under the data center on both hosts, but not if I look at events for the resource pool
Hi,
This is definitely your problem. These events are NOT normal, and this is exaclty what is filling up your database. You could try and find what (who) is logging in (from localhost so it should be something in the service console), or just reinstall to be on the safe side. All these events are logged into the VC database, I don't think they get purged at all.... I have had similar problems, using SQL queries to clean them out worked partially, but still my database was too big, so I ended up deleting it and starting over after removing the cause (in my case a 3rd party tool that produced a lot of events)
Thanks for all the responses, I found out OVPA 4.7 is the cause. Anyone else running OVPA on thier ESX hosts and seeing this issue? for now I have stopped it, and am looking for solutions.
Have you created a support request at HP ?
Are you running the latest agents specificly for vmware esx ?
We do not use hp openview performance agents on our hp servers but i hope hp will create a fix for you...