Indexes with fragmentation over 90% found during DB maintenance.

[FunkBuddha]

I saw a post about this in the RC forum but I didn't see a resolution. I'm getting this event now in the full release of 10.2.

How do I go about defragmenting?

[HemiTruck]

ok found the answer on this issue.

with the new version doing the topology tracking it seems that the data being placed in these tables is the issue

the exact tables that are causing everyone to see this now are TopologyData and NodeL2Connections and NodeL3Entries

prior to these tables NPM policed itself well and with maintenance kept the fragmentation down

now it seems all NPM databases will be getting this index fragmentation error because of the above tables.

I make this claim because all 3 of my copies of NPM are now showing this same issue and I finally beat up the DBA to find out why.

to give you an estimated timeline I upgraded to 10.2.2 on the last one on Mar 22 and it got the index issue today 2 months later.

Hope this helps

[pacetti]

FunkBuddha,

This SolarWinds Knowledge Base article should provide some insight:

"Critical index fragmentation detected during database maintenance" (http://knowledgebase.solarwinds.com/kb/questions/3202/)

HTH,

[mcparlandj]

I read over this: 

"Critical index fragmentation detected during database maintenance" 

We only run whatever sql server was installed when we installed Solar Winds Orion. It looks to be sql server 2005. 

The "SQL Server Management Studio" is not installed. 

Not sure what to do now. 

[jljeff]

Since applying this template to the SQL servers in our organization...it seems eery single server is triggering an alert on this particular value for the table spt_values.  What is the best way to fix this...is it an issue w/ the template or is there a problem on all of these sql servers?

[FunkBuddha]

Thanks for the links.

I guess my days of avoiding becoming a DBA are over. Just looking at those pages are making my eyes bleed and my mouth dry.

If NPM can detect that indexes are fragmented, could it not also report which indexes it discovered and run a job to rebuild them?

[AndyCoates]

In the past we've done this with standard SQL server commands.  Have a look at this article for more information:

http://www.simple-talk.com/sql/database-administration/defragmenting-indexes-in-sql-server-2005-and-...

[ckalinwi]

There's also this snippet from the help file:

"If you are still having issues with index fragmentation or if you want to use the defragmentation utility that is built into your SolarWinds product, submit a technical support request."

I submitted a request to get the defrag utility and the response I got back was, paraphrased, "ask your DBA to defragment the indices."  I agree with another poster in this thread - if NPM can detect which indices are fragmented, why can't it tell us and/or fix it?  Why does it require us to  learn MSSQL and hunt around to take care of this?

[harrijs]

I just started noticing these messages over the weekend.  We have been running 10.2 RC4 since it was made available.

My DBA just informed me that these are all Netflow tables and as soon as he re-indexes them more become 90%+ fragmented.

Has anyone else observed this behavior?

[netlogix]

In my experience with other databases, this could be because the "Clustered Index" is not sequential (like the pk is guid and is set to the clustered index).  And/or the primary key is the date/time the flow was generated (not flow received) and there are some flows that are getting there quicker causing a weird 'jitter' effect and messing up the index, with this one, you could decrease the 'fill factor' to allow some white space in the index for back fill.