memory leak in zabbix proxy 1.6.1

[cioris]

I installed the new 1.6.1 about 2 weeks ago (immediately after it was released). The server works fine so far, but the proxy seems to have a problem with a memory leak. I didn't have time to debug the problem, but for sure it is related to the proxy. In about 5 days it fills the entire memory (both RAM and swap). Just restarting the proxy is enogh to go back to normal... for a while, meaning another few days. As I don't have time to debug now, I just set a trigger for the swap. When it goes under 300M I automatically restart the proxy (action defined for that event).

Is anybody else having the same problem? I know there was a problem w/ a memory leak in the server 1.6 code and I thought upgrading to 1.6.1 will fix my proxy too, but it didn't. BTW, the same problem with the proxy 1.6.

Thanks

[cioris]

nobody having the same problem..??? ..

[akatsuki]

This thread will be useful for you.

http://www.zabbix.com/forum/showthread.php?t=10566

Try patching.

[cioris]

I'm using 1.6.1

The discussion is about 1.6. The new 1.6.1 is supposed to have a fix for it, but unfortunately, my 1.6.1 proxy has problems with memory leaks...

[akatsuki]

oh sorry. I overlooked the version.

[Alexei]

We are not aware of any memory leaks in 1.6.1.

[tekknokrat]

We have also memory problems with zabbix_proxy eating up all memory and taking a lot of swap right now.

10970 zabbix    21   5 2828m 2.7g 1212 S    0 90.2  34:43.75 zabbix_proxy

Could it be that proxy part was not fixed when fixing server mem-leak?

[cioris]

Any news?

any idea where the leak can come from? any tool that can be used for debugging? I don't have time now, but maybe starting w/ Dec I will have more time to dig into this problem.

Please keep me posted... I'll do the same...

[kc5kse]

I, too, am experiencing a problem with swap due to zabbix_proxy. A stop and restart fixes it for about a week.

[elvar]

I'm not sure if I am having this or not, but my Zabbix server at a colo is currently unreachable and I just recently upgraded to 1.6.1. Just before becoming unreachable I received Zabbix alerts saying that the swap file was used up.

[Alexei]

Well, the memory leak is not identified yet. What item type do you use on the proxies? Passive, active, SNMP, anything else?

[elvar]

Once I am able to get someone to go to the colo and get console access I'll be able to provide more information.

[elvar]

Well, the memory leak is not identified yet. What item type do you use on the proxies? Passive, active, SNMP, anything else?

I am not even using the proxy funtionality. I actually built Zabbix 1.6.1 without proxy support. I finally was able to the server at the colo back up and running though I'm leaving Zabbix disabled for now. The box had been up solid for 102 days running 1.4.2. Two days after upgrading to 1.6.1 I received the alerts that the box was out of swap space. There was no other known changes other than changing Zabbix versions.


Any further thoughts?

Thanks!

[elvar]

I just ran across a different post regarding memory in Zabbix and the last post mentions that the actual memory leak may be in net-snmp and not Zabbix. I checked and Ubuntu Hardy does use that version of net-snmp mentioned. I'll have to try running Zabbix but disabling snmpd and see what happens.

http://www.zabbix.com/forum/showthread.php?t=7126&page=4

[teferi]

I'm haveing thee same problem. also posted it on the bugtracker.

useing simple checks, active and passive items, althow now I intend to switch all passive to active. If the problem will disapear - will reply about it here