The kernel hardware watchdog is supposed to handle restarts even when userspace processes are starved out from the kernel. Watchdog runs with real time priority and if the kernel detects the watchdog is falling being the system is rebooted. I would suggest to reconfigure, not disable. I do not think saying ha is a replacement for the kernel watchdog.
On Mon, Aug 25, 2008 at 1:50 PM, Lars Ellenberg <[EMAIL PROTECTED]> wrote: > On Mon, Aug 25, 2008 at 12:52:28PM -0400, Doug Eubanks wrote: >> I'm pretty sure that's not the case. It's immediate as soon as the server >> hits a 1min load average of 30. It does not matter how long it took the >> server to reach 30. >> >> We are using heartbeat 2.1.3. >> >> I saw this message today.... >> >> Aug 25 12:49:14 mailserver1A watchdog[2316]: loadavg 25 11 4 is higher than >> the given threshold 24 18 12! >> >> Think it may be related? > > sure that is related. > but not to heartbeat. > > man watchdog > man watchdog.conf > vi /etc/watchdog.conf > > either configure that watchdog thing > "correctly" for your needs and setup. > > or get rid of it completely. > after all, by using heartbeat, > you already have your cluster nodes "monitoring" each other. > and the watchdog ain't no use if it gets in the way, right? > > > -- > : Lars Ellenberg > : LINBIT HA-Solutions GmbH > : DRBD(R)/HA support and consulting http://www.linbit.com > > DRBD(R) and LINBIT(R) are registered trademarks > of LINBIT Information Technologies GmbH > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
