1.6.3 very High Load

[753713]

Hi,

ater upgrade from 1.6.2 to 1.6.3 I have very high Load on my Zabbix Server ~7-9, and my Windows Eventlog Monitoring is no running on all Windows Server.

After downgrade is my Load ~2-3 and the Zabbix Server running perfectly without errors.

any ideas for my problem with the version 1.6.3

thanks
753713

[bbrendon]

What parts of zabbix did you upgrade/downgrade?

[Alexei]

This is hard to believe. Zabbix 1.6.3 introduced some features which makes it run faster, not slower!

[753713]

Hi,
with zabbix_server 1.6.2 running my server perfektly, after change zabbix_server to 1.6.3 running my server with load ~7-9.

Frontend replace has no impacts os server load.

[Alexei]

Anyone else having similar issues?

[bimind]

I have the same problem.

After the update from 1.6.2 to 1.6.3 CPU load and CPU utilization have increased in my zabbix

I updated the binaries and frontend.

Thanks,
Oriol

[bimind]

The CPU graph after upgrade

[753713]

in the version 1.6.4 is the load also very high

[jmcnulty]

New indexes?

Have those of you experiencing problems here created the indexes advised on the Zabbix 1.6.4 information page?

I've upgraded but have not done this yet as I'm not sure of the exact steps required.

John

[bimind]

The problem appears with and without the indexes.

Oriol

[t2y]

Hi Oriol,

The problem appears with and without the indexes.

I'm interested in this topic. Do you mean that we should create as follows indexes which were written by Release Note?

CREATE UNIQUE INDEX history_log_2 on history_log (itemid,id);
CREATE UNIQUE INDEX history_text_2 on history_text (itemid,id);
CREATE INDEX graphs_items_2 on graphs_items (graphid);

[753713]

I have install upgrade to 1.6.4 and the index is create.
1. with 1.6.3 I have very high load ~7-9 and my Windows Eventlog Monitoring is not running.

2. with 1.6.4 I have high load ~5-7 and my Windows Eventlog Monitoring running perfectly.

Any Ideas how can I the load smaller configure?

[bimind]

Hi everybody,

Any news about this thread? I continued with the problem!

Thanks,
Oriol

[Kerrygeek]

Hi Oriol,


I'm interested in this topic. Do you mean that we should create as follows indexes which were written by Release Note?

CREATE UNIQUE INDEX history_log_2 on history_log (itemid,id);
CREATE UNIQUE INDEX history_text_2 on history_text (itemid,id);
CREATE INDEX graphs_items_2 on graphs_items (graphid);

I just upgraded my test box from 1.6.1 to 1.6.4 hoping to get away from the memory leak problem. Is there any reason I can't go ahead and make these indexes on my 1.6.2 production server before I do the upgrade?

Thanks,
Kerry

[Kerrygeek]

I upgraded my Zabbix server over the weekend and also built the indexes listed earlier in this thread. My utilization seems to be down, I'll have to let it run a while to determine if the memory leak is fixed but so far it looks fine. I had a cron job set up to restart Zabbix every morning, I've disabled it so we'll see if it still uses up all my memory.

And, for those of you who posted upgrade instructions, thanks! It was a big help and should be included in the next version of the manual.

Kerry