1.4: agent and server still fail to properly restart

[abi]

hi,

the server/agent still fail to be restarted from the init scripts:


20147:20070531:084622 One child process died. Exiting ...
20166:20070531:084622 zabbix_agentd started. ZABBIX 1.4.
20166:20070531:084622 Listener failed with error: Cannot bind to port 10050 for server [ANY]. Error [Address already in use]. Another zabbix_agentd already running ?.
20147:20070531:084624 ZABBIX Agent stopped


seems like its still the same old REUSE ADDR issue, just that its now failing
on *each* erstart, not occassionally .. :(

[abi]

it helps adding a sleep statement (about 2 seconds) to the init scripts..
not nice but ... well :)

[juampe]

If you are using external scripts, kill all external scritps before start again

[Farzad FARID]

it helps adding a sleep statement (about 2 seconds) to the init scripts..
not nice but ... well :)

Hi,

That's what I did in my modified version of the Fedora Core init scripts.. Not nice but I know many packages that need this kind of delay.

So unless the source code is modified upstream to reuse the socket more efficiently, I second the suggestion of adding "sleep" to the init script.

Regards

[abi]

Hi,

That's what I did in my modified version of the Fedora Core init scripts.. Not nice but I know many packages that need this kind of delay.

So unless the source code is modified upstream to reuse the socket more efficiently, I second the suggestion of adding "sleep" to the init script.

Regards

yeah well, there are several other init scripts in debian that do the same
thing. A sleep statement during restart isnt *that* bad, as it doesnt slow
down the bootup etc.

abi@ke:~$ grep sleep /etc/init.d/* | grep -c sleep
35

seems to be common practice, ..

[fatiha]

Hi,

Today I has the same problem, so I stopped zabbix_agentd and I killed external scripts and now it works ! So thank you juampe.

Fatiha