I have a problem which has only just started happening after no obvious server changes. We are using networker 18.104.22.168. (Windows 2008 R2)
every night when backup jobs run it seems that NSRIM.EXE is using 30GB of memory. This causes jobs to fail and Networker to crash, mainly because of timeout error and resource issues because of this process.
I can get back into the networker GUI by shutting down all services and processes, this has happened now for 2 weeks and not occurred before.
EMC support saying we should upgrade to 22.214.171.124 - There release notes do mention a bug for a memory leak but is specific to Exchange only. I have no idea why this bug has only just started happening, I am hoping someone can shed some light on this and perhaps a workaround to the problem other then upgrading.
There are various reasons as of why could be happening, such as lag in the expiration of savesets, long queue of savesets to be deleted due to devices being busy etc.
I would strongly recommend to upgrade to latest built, as you are just using the first cumulative, as there has been multiple code improvements and fixes related to nsrmmdbd, which is the process with whome nsrim interacts directly, including some memory leaks as well, so first step I would recommend would be that, upgrading NW server and storage nodes to 126.96.36.199
I second what support said. As you use SP3 it makes sense to use latest post SP3 patch which is 188.8.131.52. On Linux I haven't seen issues yet in that domain so this could be Windows specific, but I suspect 184.108.40.206 could really fix it for you.
This is the memory leak we had last night which has been typical, with what you are saying would you say the spike is caused by jobs ending at this time?
Or should I be looking at jobs starting? The time is between 430-5AM that it happens, it drops when I get to work and force NSRIM to quit.
Hello, we get kind the same behavior in the past. As workaround, we have renamed NSRIM.EXE into NSRIMABC.EXE, and scheduled it with Microsoft scheduler. It guaranteed that the nsrim process is running outside our backup windows and not 12 hours after networker service start.
We have to do this renaming by every networker update.
Hope it will help you.
Rather than renaming nsrimexe, would be better to "touch" nsrim.prv, and then schedule it with Windows task scheduler.
If you schedule it but don't rename the .exe file, it will never run, as it won't find it due to the name change.