is this normal: iwserver running on 1GB of memory?

Hello,
i am on TS 5.5.2 on w2k... and sometimes our teamsite goes very slow and have to reboot the box... the last time i rebooted it iwserver was running on 1GB of memory of 1.5GB that all the process together were using and we have 2GB available for all processes.

we have about 19 teamsite user and the iw-store is 12gb.
when i rebooted the box it start on 100MB and not it s on 850MB and i affraid the box will go slow again... is the memory usage of iwserver normal for our case?

the iw-home/local/logs dont tell me much about the issue... what r the problem logs to check why our teamsite goes slow too slow every few days?

Find more posts tagged with

Comments

pdasarat

Do you have some custom login modules in place?. We had this issue where we saw a huge peek increase when user login and it never used to drop down. We found that there were a memory leak with the custom apache module we created. Hope it helps.

-Pradeep

CRB

that sounds very close to the issue we have here, but i dont think we have any custom module... how did you fix the issue?

pdasarat

Believe me or not, it took us a long time (about 4 months) to get to the root problem. No erros, no warnings, nothing special in the logs. everything was perfect, expect the process tends to grow fast.
In our case we had siteminder integration using a custom written apache module. We later found that it was due to an apache API change, forgot the name. We re-build the module and it was back to normal. Atleast there was change in pattern of memory increase. Hope it gives you some indicator.

-Pradeep

CRB

yeah, i have been watching all day, it grew from 100mb to 915mb right now, and it never went down...
before u got to fix it (for 4 months) what did you do to get around it? for now, we just wait for calls from the user and then reboot... and that is not efficient way... did you use something to monitor and warn you? or what did you do about it?

pdasarat

We did a nightly refresh of the servers (15 mins downtime) to keep the support incidents low. We had users across the globe, but we communicated to the business to accept this temporary solution till resolved.