[ClusterLabs] Antw: [EXT] Re: VirtualDomain & "deeper" monitors - what/how?

[Kyle O'Donnell]

I am thinking about using the qemu-guest-agent to run one of the available commands to determine the health of the OS inside

virsh qemu-agent-command myvm --pretty '{"execute":"guest-get-fsinfo"}'

https://qemu-project.gitlab.io/qemu/interop/qemu-ga-ref.html

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐

On Thursday, May 13th, 2021 at 01:28, Andrei Borzenkov <arvidjaar at gmail.com> wrote:

> On 03.05.2021 09:48, Ulrich Windl wrote:
>
> > > > > Ken Gaillot kgaillot at redhat.com schrieb am 30.04.2021 um 16:57 in
> > > > >
> > > > > Nachricht
> > > > >
> > > > > 3acef4bc31923fb019619c713300444c2dcd354a.camel at redhat.com:
> > > > >
> > > > > On Fri, 2021‑04‑30 at 11:00 +0100, lejeczek wrote:
> > >
> > > > Hi guys
> > > >
> > > > I'd like to ask around for thoughts & suggestions on any
> > > >
> > > > semi/official ways to monitor VirtualDomain.
> > > >
> > > > Something beyond what included RA does ‑ such as actual
> > > >
> > > > health testing of and communication with VM's OS.
> > > >
> > > > many thanks, L.
> > >
> > > This use case led to a Pacemaker feature many moons ago ...
> > >
> > > Pacemaker supports nagios plug‑ins as a resource type (e.g.
> > >
> > > nagios:check_apache_status). These are service checks usually used with
> > >
> > > monitoring software such as nagios, icinga, etc.
> > >
> > > If the service being monitored is inside a VirtualDomain, named vm1 for
> > >
> > > example, you can configure the nagios resource with the resource meta‑
> > >
> > > attribute container="vm1". If the nagios check fails, Pacemaker will
> > >
> > > restart vm1.
> >
> > "check fails" mans WARNING, CRITICAL, or UNKNOWN? ;-)
>
> switch (rc) {
>
> case NAGIOS_STATE_OK:
>
> return PCMK_OCF_OK;
>
> case NAGIOS_INSUFFICIENT_PRIV:
>
> return PCMK_OCF_INSUFFICIENT_PRIV;
>
> case NAGIOS_NOT_INSTALLED:
>
> return PCMK_OCF_NOT_INSTALLED;
>
> case NAGIOS_STATE_WARNING:
>
> case NAGIOS_STATE_CRITICAL:
>
> case NAGIOS_STATE_UNKNOWN:
>
> case NAGIOS_STATE_DEPENDENT:
>
> default:
>
> return PCMK_OCF_UNKNOWN_ERROR;
>
> }
>
> return PCMK_OCF_UNKNOWN_ERROR;
>
> Manage your subscription:
>
> https://lists.clusterlabs.org/mailman/listinfo/users
>
> ClusterLabs home: https://www.clusterlabs.org/