Hi to everyone.
I have some strange problem with Unreachable poller at SNMP items. If some of my hosts starting to response longer than usual (for example becouse of the network load), my Unreachable poller raising severely (the busy status). But even after returning to normal the response time problem is still here. Unreachable poller is under heavy load and the items still staing in "more than 10 min queue" but the connection is Ok.
I have a filling that the Unreacheble poller don't using timeout parameter at all. It's starting checking availability of the item and waiting forever for responce. And becouse of this it's not going for checking next item. But how to check it?
If I'm rebooting zabbix server it's returning to normal until first problem with connection.
This is how the Unreachable poller is raising:
I reloading zabbix server until poller reach 100% becouse in this way my zabbix server stoping collecting any SNMP information.
Additional information:
All items is SNMPv3 agent. There is no such problems with simple checks or zabbix agent.
OS:
#uname -a
Linux Zabbix 2.6.32-431.3.1.el6.x86_64 #1 SMP Fri Jan 3 21:39:27 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
#cat /etc/issue
CentOS release 6.5 (Final)
Zabbix (with MySQL):
Arch : x86_64
Version : 2.2.1
Release : 1.el6
Config file:
StartPollers=60
StartPollersUnreachable=60 #Such a high parameter just for the test (but result at the graph)
StartPingers=20
StartDiscoverers=1
StartTimers=5
MaxHousekeeperDelete=3000
SenderFrequency=60
CacheSize=64M
StartDBSyncers=6
HistoryCacheSize=16M
ValueCacheSize=24M
Timeout=3
# UnreachablePeriod=60 # Deafult value
UnavailableDelay=30
UnreachableDelay=4
LogSlowQueries=1500
I have some strange problem with Unreachable poller at SNMP items. If some of my hosts starting to response longer than usual (for example becouse of the network load), my Unreachable poller raising severely (the busy status). But even after returning to normal the response time problem is still here. Unreachable poller is under heavy load and the items still staing in "more than 10 min queue" but the connection is Ok.
I have a filling that the Unreacheble poller don't using timeout parameter at all. It's starting checking availability of the item and waiting forever for responce. And becouse of this it's not going for checking next item. But how to check it?
If I'm rebooting zabbix server it's returning to normal until first problem with connection.
This is how the Unreachable poller is raising:
I reloading zabbix server until poller reach 100% becouse in this way my zabbix server stoping collecting any SNMP information.
Additional information:
All items is SNMPv3 agent. There is no such problems with simple checks or zabbix agent.
OS:
#uname -a
Linux Zabbix 2.6.32-431.3.1.el6.x86_64 #1 SMP Fri Jan 3 21:39:27 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
#cat /etc/issue
CentOS release 6.5 (Final)
Zabbix (with MySQL):
Arch : x86_64
Version : 2.2.1
Release : 1.el6
Config file:
StartPollers=60
StartPollersUnreachable=60 #Such a high parameter just for the test (but result at the graph)
StartPingers=20
StartDiscoverers=1
StartTimers=5
MaxHousekeeperDelete=3000
SenderFrequency=60
CacheSize=64M
StartDBSyncers=6
HistoryCacheSize=16M
ValueCacheSize=24M
Timeout=3
# UnreachablePeriod=60 # Deafult value
UnavailableDelay=30
UnreachableDelay=4
LogSlowQueries=1500
Comment