Forum Discussion
Nimbostratustroubleshooting slow performance
folks
i'm getting poor performance for my proxy environment and i'm hoping you can help
i have an ltm 1600 (11.4.1) cluster in front of 3 ironports (C670s) but performance is poor when more and more users start to use the service
in the morning and late afternoon things are very good but the rest of the day browsing is very poor
if i check the ltm with the show ltm persistence persistence-records things run well up to 3000 records, after that a web page can take 30 secs to load
i'm using a http amd https profile with a different vip for each one, there is no decryption taking place and a simple irule to log
the ltm shows 60% utilisation on the ltm and 18% on the ironport
can anyone give an indication of how to properly troubleshoot the ltm to make sure its not the cause of poor performance
thanks
4 Replies
What_Lies_Bene1Cirrostratus
Can you describe how the VS is setup please? Standard? Performance L4? What Persistence method, any iRules etc? 3k records on an LTM should be no problem at all, persistence records consume RAM, not CPU in the main so that 60% is probably due to some poor configuration.
- What_Lies_Bene1
You're welcome.
What's the iRule logging? I suspect this is the issue (unless you're using HSL). Can you post it?
Is a HTTP profile assigned?
Standard TCP profile?
- nitass
Employee
i would start by checking tcpdump to see where the slow is.
e.g.
tcdpump -nni 0.0:nnn -s0 -w /var/tmp/if snat is not used, you may filter by using client ip (host x.x.x.x).
- What_Lies_Bene1
I'd agree with Nitass but I'd also make the point here that with that iRule you are logging a huge amount of data surely? If you've 3k clients and you are logging on every HTTP request using the syslog-ng program rather than HSL that could be the cause, that's a log of messages! Can you try removing the rule and seeing if things improve.
