[Check_mk (english)] Unresolvable race condition in 'ntp' check code.
mk at mathias-kettner.de
Wed Feb 22 18:15:05 CET 2012
I've now removed the according condition from the check. I do not think
that it is very important anyway.
Am 25.11.2011 19:56, schrieb Stier, Matthew:
> Part of the ‘ntp’ check code (/usr/share/check_mk/checks/ntp) tests to
> see if the seconds since last poll results (when) is less than 2 times
> the current poll interval (poll).
> A race condition exists when there is a transition from a high to low
> ‘poll’ interval, a high ‘when’ count created by the previous high poll
> What happens is that while the ‘poll’ interval is reduced immediately,
> there may be a delay to when the next poll is initiated, it’s response,
> and the resetting of the ‘when’ counter; in that interval Nagios can run
> an NTP check and will get a false failure. The follow up check by Nagios
> one minute later clears the failure, but, the false failure will already
> have been logged, and fault/recovery notifications sent.
> This happens with my three primary NTP servers (with their clock sources
> across a 70ms round trip WAN link) three to four times per day. My only
> solution so far, is to comment out the test condition, and it’s single
> line conditional block of code.
> checkmk-en mailing list
> checkmk-en at lists.mathias-kettner.de
Mathias Kettner GmbH
Registergericht: Amtsgericht München, HRB 165902
Firmensitz: Preysingstraße 74, 81667 München
Geschäftsführer: Mathias Kettner
Tel. 089 / 1890 4210
Fax 089 / 1890 4211
More information about the checkmk-en