System Uptime

[kbldawg]

I've noticed since switching out all my old WRAPs to WARs that my system uptime is very low, in comparision to the wraps, on all systems.

Otherwise they all seem to be working very well.

So...I am curious if anyone else is seeing the same thing I am.

[tog]

I'm not sure I understand you... you have had WRAPs running for a long time and your newly-installed WARs have a shorter uptime?

Are you trying to say indirectly that you are having problems with WARs spontaneously rebooting/crashing when you were not having any such problems with the WRAPs?

My WRAP-based APs running v2 were considerably less stable than my WAR board APs running v3.

[kbldawg]

exactly...

I have 3 WRAPs left on my network. Of those 3, 2 are in the hundreds of days. This was typical for the WRAP boards, in my experience.

Out of 20 or so WAR4s I do not have one single board with 30 days or more uptime; as a matter of fact...typically they are less than a week on average.

screenshot of what I'm talking about...

http://lists.tcworks.net/~chalsted/Junk/uptime.bmp

The left side shows WAR4 boards and their uptime. The right side shows the 3 remaining WRAPs on my network.

I did not pick and choose the WARs, they were the first in my putty list.

All the WRAPs that were replaced with WARs are using the same ethernet cable, jacks, patch cables, poe surge...etc. The only thing different at these sites is the board, ps and any prisms got swapped out with CM9s and WLM54Gs. The AP configs are even the same.

Once SNMP support was added we did start polling those for MRTG data. Not sure if this could be an SNMP thing or not.

[tony]

SNMP has a known (and every elusive) memory leak that will cause the system to reboot after several days. While this will not adversely effect your network stability & performance, it will keep the uptime counters from giving you those 300+ days uptime.

[amp]

Got 2 war boards with uptime. Doesn't seem like the 1.0.1 war board lasted past 150 but thats hard to say. So war boards at least don't reboot.

http://img96.imageshack.us/img96/6140/waruptimesiy8.jpg

[butchkemper]

Is there any update on the StarOS V3 problem with SNMP and possible solution?

Butch

[tony]

No solutions yet. It's an internal issue with net-snmp that we are investigating.

[kbldawg]

Thanks Tony, as I said, the systems run great other than the reboots.

[tog]

Hehe you should turn SNMP off.

I have seen the SNMP leak cause a loss of ability to bring up the ssh GUI and also once loss of ability to even do a starutil -reboot, required a power cycle.

I really wouldn't recommend leaving SNMP on for any reason, it's kind of dangerous to do so.

[lonnie]

We have SNMP turned off until the source of the leak is found. I advise others to do the same.

[kbldawg]

Thanks everybody, sound advice and I have decided to take it. SNMP is off until a fix.

Tog,

I've definately seen the login problem, very frustrating. After you enter your password, some units, at random times, will just hang up. Sometimes, if you wait long enough, they go ahead and log in, other times I've had to just kill the session and try again and other times the only way I could get back in was to do a starutil -reboot and then try it right after it came back up.

[kbldawg]

Just an update...

Since disabling SNMP all random reboot issues have gone away.

System Uptimes have improved and are growing everyday.