Forum Discussion
djfcc_23029
Nimbostratus
NimbostratusMonitoring Flapping
Hi,
I need some help investigating a problem with our GTMs where by all pools at one of our two sites consistently changes state between available and down.
Our set up is two GTMs in one sync group, at two different data centres. We use the GTMs to load balance a simple HTTP web service between the two sites.
Our monitoring is configured to use HTTP monitors to make a simple HTTP request and check for a response string. The HTTP monitor timeout is set to 121 seconds, the interval at 30 seconds and the probe timeout of 5 seconds.
The monitors report that the pool members are timing out but I know this is not the case. We monitor all of our services from three different locations directly (Not via the GTMs) and all of the services respond generally in sub second response times but occasionally very small blips of up to 3 seconds.
Any ideas how I can get more information(debugging information) out of the GTMs to help find the solution would be great? Any ideas what might be the cause would be even better!
Thanks,
David.
hooleylistCirrostratus
Hi David,
Have you tried running tcpdump on the GTM which is monitoring these services? Does the failure occur often enough to do this?
Aaron
vaibhavtry to have a tcpdump/iqdump when you are having this issue ....also validate the logs once u see them failing......it also happenes when thr are some network hindrance occurs at site ........
djfcc_23029Hey Guys,
Thanks for your replies sorry I've not replied sooner. The tcpdump did help verify that the time outs we were seeing were genuine. After alot of network dumps and traces I managed to identify a dell switch as the cause of the problem.
Thanks, for all your help.
David.- hooleylistGlad to hear. Thanks for posting the resolution.
Aaron 
AneshCan you tell me what was wrong with the switch??