NPM 12.2 port exhaustion - Fixed by 12.4?

Just trying to look before I leap, as it were, since I'm not any kind of SW expert by a long shot...

We're currently running NPM 12.2.5300.360. which we upgraded to last spring. Completely new VM environment - new servers, new OS, etc.

Since then we've seen repeating (but not entirely consistent) issues with what turned out to be port exhaustion. The app server / poller will slow down to the point of eventually requiring a reboot, usually from the VM console. A support case ensued which exposed no obvious culprit until finally the port exhaustion finding. As a temporary workaround we maxed out MaxUserPorts and it, along with additional RAM and vCPUs, has definitely helped. But the slowdowns still come and go.

I was told by the SW support engineer that the issue was fixed by 12.3 HF6, which he also said didn't touch the database - Just an app patch only. When installing it, it became clear it was a full upgrade so I halted and consulted SW Support again. That engineer confirmed and also indicated that it did not, in fact, fully fix the port exhaustion issue. He said the problem was still being worked and I might want to wait for the next release. So, I waited and now 12.4 is out.

Does anyone know if 12.4 actually fixes port exhaustion? I'd hate to go through an upgrade only to have the problem go unfixed. This is particularly true since we have no test environment.

Thanks.

Find more posts tagged with

Accepted answers

All comments

neomatrix1217

This was addressed in Orion Core Hotfix 2018.2. HF4

Orion Platform 2018.2 Hotfix 4 - SolarWinds Worldwide, LLC. Help and Support

mikeit83

I am 100% current on our production system. No port exhaustion issues- and our previous issues of random poller status down issue appears to have gone away.

We had quite a few support cases that appear to have been self inflicted by trying to do the HA upgrade to convert from 2012 to 2016. If I would have skipped that, I wouldnt have had strange poller entries in the database causing my grief.

Been stable for 2 weeks with no need to reboot or recycle services. Its been super fast and generally....awesome. I can actually do work now instead of following around the orion servers with defibrillators every waking moment

lphilpot

Sounds reassuring. I'll have to check it out and hopefully schedule an upgrade. While I spend little time in SW, others use it pretty much all day. Need to keep them happy. :-)

Thanks for the info.

mikeit83

Just make sure you have good backups and understand the pre-req's and system requirements before you jump. And most importantly, good backups of every server (especially database). Yes I said backups twice. its that important.

Dont forget the change controls and coffee.

Ohh- and most importantly 1866-530-8040 option 3. I have it on speed dial.

lphilpot

OS backups and SQL backups will be verified, plus VM snapshots of all servers before I start. And testers lined up in advance. :-) I already have SW Support's number on my whiteboard. We went through a Two. Week. Long. upgrade last spring to implement this environment (partly with a consultant and partly by myself). I have no interest in doing that again...

Thanks.