web checks fail "doing curl_easy_perform: Timeout was reached", but after restartin z

[sean]

Hi,

I was getting "doing curl_easy_perform: Timeout was reached" for a webcheck, but on the command line with curl it worked fine, taking about 2 secs.

Finally I restarted the zabbix server daemon, and now its fine.
v2.0.4

So it seems like a bug in the server to me.
Here more longs if it helps:
26172:20130503:114823.658 sleeping for 1 seconds
26192:20130503:114823.688 web scenario step "foo: bar" error: error doing curl_easy_perform: Timeout was reached
26192:20130503:114823.688 In process_step_data() rspcode:0 time:0.000000 speed:0.000000
26192:20130503:114823.688 query [txnlev:0] [select type,itemid from httpstepitem where httpstepid=57]
26192:20130503:114823.689 End of process_step_data()
26192:20130503:114823.689 query without transaction detected
26192:20130503:114823.689 query [txnlev:0] [update httptest set nextcheck=1367574503+delay where httptestid=30]
26170:20130503:114823.808 In get_pinger_hosts()
26170:20130503:114823.808 In DCconfig_get_poller_items() poller_type:3
26170:20130503:114823.808 End of DCconfig_get_poller_items():0
26170:20130503:114823.808 End of get_pinger_hosts():0
26170:20130503:114823.808 In process_pinger_hosts()
26170:20130503:114823.808 End of process_pinger_hosts()
26170:20130503:114823.808 icmp pinger #1 spent 0.000374 seconds while processing 0 items
26170:20130503:114823.808 In DCconfig_get_poller_nextcheck() poller_type:3
26170:20130503:114823.808 End of DCconfig_get_poller_nextcheck():1367574504
26170:20130503:114823.809 sleeping for 1 seconds
26214:20130503:114823.848 In collect_selfmon_stats()
26214:20130503:114823.848 End of collect_selfmon_stats()


Also a little question: when a webcheck is assigned to a client, it still runs on the server? The assignment makes no difference?

Regards,

Sean Boran

[Heilig]

Also a little question: when a webcheck is assigned to a client, it still runs on the server?

Yes. Zabbix server does not use agent for web checks.

The assignment makes no difference?

Has thus you can identify the problem with the correct host (highlight it, send a notification to right people, etc).