-
Incident report
-
Resolution: Fixed
-
Trivial
-
4.0.12
-
None
-
Sprint 56 (Sep 2019), Sprint 57 (Oct 2019)
-
0.5
On Suse 11.4 while using 4.0 LTS agent with incorrect 'ListenIP=' configured each agent attempt (not successful start) will allocate a semaphore:
------ Semaphore Arrays -------- key semid owner perms nsems 0x00000000 65536 zabbix 600 14
Each next attempt will allocate even more semaphores:
------ Semaphore Arrays -------- key semid owner perms nsems 0x00000000 65536 zabbix 600 14 0x00000000 98305 zabbix 600 14 0x00000000 131074 zabbix 600 14 0x00000000 163843 zabbix 600 14 0x00000000 196612 zabbix 600 14 0x00000000 229381 zabbix 600 14 0x00000000 262150 zabbix 600 14 0x00000000 294919 zabbix 600 14 0x00000000 327688 zabbix 600 14 0x00000000 360457 zabbix 600 14 0x00000000 393226 zabbix 600 14 0x00000000 425995 zabbix 600 14 0x00000000 458764 zabbix 600 14
Steps to reproduce
Download zabbix-agent-4.0.12-1ent.sles11.x86_64.rpm from https://enterprise.zabbix.co.jp/update/12166 (package not available publicly)
install package:
rpm -i zabbix-agent-4.0.12-1ent.sles11.x86_64.rpm
Set the incorrect 'ListenIP' directive in /etc/zabbix/zabbix_agentd.conf:
ListenIP=192.126.2.99
Start agent:
zabbix_agentd -c /etc/zabbix/zabbix_agentd.conf -f # agent will not start
Check semaphores:
ipcs -s
P.S. such behaviour does not happen with Zabbix agent 4.0.12 package on CentOS 7.6