Re: Monit state is not fully saved

[Martin Pala]

Hi Karen,

that problem is fixed in the development version already, snip from changelog:

--8<--
Fixed: Issue #316: The "if changed checksum" and "if changed timestamp" tests 
value
is persistent across monit restart/reload now, so if the checksum changed while 
monit
was stopped or reloading, it will catch the change. Thanks to Duke Bartholomew 
for fix.


Fixed: Save the file size, filesystem flags, file/directory/fifo/filesystem 
permissions,
network link speed so the "if changed" tests keep the last value across monit 
restart/reload.
--8<--

Best regards,
Martin


> On 13 Feb 2016, at 12:48, karen.arutyunov <address@hidden> wrote:
> 
> Hi,
> 
> I have just started to evaluate Monit 5.16 and have noticed the following 
> behaviors.
> 
> 1. File timestamps are not saved to .monit.state file, so CHECK FILE 
> statement misses the file timestamp change if it happen between Monit 
> executions in daemon mode. This somewhat inconsistent with the fact that the 
> file read position IS saved, so Monit being configured to monitor log file 
> for error messages can detect those which appeared between Monit executions.
> 
> 2. If executed in non-daemon mode Monit reads the state from .monit.state 
> file, but do not save the latest state on exit (unlike the daemon mode). This 
> make it impossible to use non-daemon mode to monitor logs for specific 
> messages.
> 
> Would very much appreciate if someone explain if these behaviors are for 
> reason, or bugs, or well known but planned to be changed.
> 
> Best regards,
> Karen
> 
> 
> --
> To unsubscribe:
> https://lists.nongnu.org/mailman/listinfo/monit-general