Hi, as mentioned, you can use the generic send/expect test: http://mmonit.com/monit/documentation/monit.html#GENERIC-SEND-EXPECT
I can implement native sieve protocol test too ... it would help if you can provide access to SIEVE's TCP port, so i can save some time. Regards, Martin > On 11 Nov 2014, at 13:25, Jorge Bastos <[email protected]> wrote: > > Hi Martin, > > I've confirmed this, when it gets zombie, monit doesn't restart it. > The TCP port stays open, but no response is obtained. > > What can I do to help you on this? > >> -----Original Message----- >> From: [email protected] >> [mailto:[email protected]] On >> Behalf Of Martin Pala >> Sent: quinta-feira, 6 de Novembro de 2014 21:59 >> To: This is the general mailing list for monit >> Subject: Re: Protocol for a SIEVED ? >> >> If the port test succeeded (no errors), then it seems that the port >> 2000 is handled by some other process (not sieved), or that the sieved >> was able to accept TCP connection and it failed later on protocol >> level. Please can you check with which process listens on this port?: >> >> sudo lsof +c 0 -P -iTCP -sTCP:LISTEN | grep 2000 # the PID will >> be in the second column >> >> vs. the PID of the sieved process: >> >> cat /rundbmail/dbmail-timsieved.pid >> >> If it accepted TCP connection but failed to handle requests, you'll >> need to use more specific protocol test ... you can use generic >> send/expect (http://mmonit.com/monit/documentation/monit.html#GENERIC- >> SEND-EXPECT) or we can implement native test. >> >> >> >> >>> On 06 Nov 2014, at 22:40, Jorge Bastos <[email protected]> >> wrote: >>> >>> Martin, >>> >>> There’s nothing on the logs, only what I show below but because I >> killed the process manually, >>> >>> check process dbmail-timsieved with pidfile /rundbmail/dbmail- >> timsieved.pid >>> every 1 cycles >>> group dbmail >>> start program = "/etc/init.d/dbmail-timsieved start" >>> stop program = "/etc/init.d/dbmail-timsieved stop" >>> if failed port 2000 type tcp with timeout 20 seconds within 1 >> cycles then restart >>> >>> >>> [WET Nov 6 21:02:34] error : 'dbmail-timsieved' process is not >> running >>> [WET Nov 6 21:02:35] info : 'dbmail-timsieved' trying to restart >>> [WET Nov 6 21:02:35] info : 'dbmail-timsieved' start: >> /etc/init.d/dbmail-timsieved >>> [WET Nov 6 21:03:35] info : 'dbmail-timsieved' process is >> running with pid 31151 >>> >>> From: [email protected] >> [mailto:[email protected]] On >> Behalf Of Martin Pala >>> Sent: quinta-feira, 6 de Novembro de 2014 21:16 >>> To: This is the general mailing list for monit >>> Subject: Re: Protocol for a SIEVED ? >>> >>> Hi, >>> >>> please can you post snip of the full service check + related monit >> log? >>> >>> Regards, >>> Martin >>> >>> >>>> On 06 Nov 2014, at 21:54, Jorge Bastos <[email protected]> >> wrote: >>>> >>>> Hi, >>>> >>>> Which protocol should i use for a SIEVED server? >>>> I’m using it with: >>>> >>>> if failed port 2000 type tcp with timeout 20 seconds within 1 cycles >> then restart >>>> >>>> But it got zombie and monit didn’t restarted it. >>>> >>>> Thanks, >>>> -- >>>> To unsubscribe: >>>> https://lists.nongnu.org/mailman/listinfo/monit-general >>> >>> -- >>> To unsubscribe: >>> https://lists.nongnu.org/mailman/listinfo/monit-general >> >> >> -- >> To unsubscribe: >> https://lists.nongnu.org/mailman/listinfo/monit-general > > > -- > To unsubscribe: > https://lists.nongnu.org/mailman/listinfo/monit-general -- To unsubscribe: https://lists.nongnu.org/mailman/listinfo/monit-general
