Rebooting VCS-C/VCS-E on occasion for Maintenance?

[Douglas Baggett]

We had an issue where a remote user on the Internet who was using STUN/TURN with ICE using Movi/Jabber could not hear or see another user behind our firewall.

Rebooting our VCS-C/VCS-E fixed the problem (in retrospect I should of just rebooted the expressway first to see if it was just the expressway).

Given that somehow the VCS-C or VCS-E got mucked up somehow, what is common practice when it comes to rebooting/re-starting them? Is it common practice to only reboot them when there are problems, or is it a good idea to cold boot them every so often, say on a weekend to clear any problems.

[Martin Koch]

In general it should not be a problem, but sure it might depend on the software version, think x6 had some

problems when you got an uptime of a year or so, ...

Had been at a customer recently which had a X4.x pair of VCS, MCU and was using it which had an

uptime of >900 days, so at least I would not generalize it.

In general I would also say thats what I would expect by a device like the VCS, it should run stable,

though maybe not the 900 days as this also means, 900 days without new features and most imporant

bug and security fixex ;-)

Which software version do you run? Do you see any other problems?

I would check for example if your firewall has the right port openings, with X7.x some new ports for

RTP (traversal and b2bua) were added, so this can cause no video in some scenarios.

I did not test if the VCS allocates the ports randomly or increasing, especially the latter could

nicely explain what you see.

Also if a call hangs and floods otherwise used RTP ports can cause strangeness.

[Douglas Baggett]

We are running 7.1 on both the VCS-E and VCS-C. The firewall seems to be set fine (since it worked after reboot). I'm wondering if your last sentence on call hangs and floods used RTP ports might have been the issue?

It's all just speculation at this point

I rebooted both less than a month ago so it had not been up THAT long.

Given both really are LINUX boxes (I have a background in Unix/Linux sysadmin) I'm wondering if there might have been a race condition in a process or set of resources that did not get freed up after previous calls. I noticed after reboot that calls connected quicker than before the boot.

If it happens again and there is not an immidiate need I'll log in and take a look at resource allocation. Couldn't do it the other day because the problem had to be solved immidiately.

thanks for the info!