[ClusterLabs] Alert notes
Ferenc Wágner
wferi at niif.hu
Wed Jun 15 16:11:55 UTC 2016
Hi,
Please find some random notes about my adventures testing the new alert
system.
The first alert example in the documentation has no recipient:
<alert id="my-alert" path="/path/to/my-script.sh" />
In the example above, the cluster will call my-script.sh for each
event.
while the next section starts as:
Each alert may be configured with one or more recipients. The cluster
will call the agent separately for each recipient.
and the rest of the documentation considers the recipient always
present. For example, in table 7.2:
CRM_alert_recipient The configured recipient
then
Alert agents will be called once per recipient.
While in specialized cases it certainly makes sense that some alerts
don't take recipients, I find it confusing that the first introductory
example demonstrates something totally unacknowledged by the definitive
text following it.
I think the default timestamp should contain date and time zone
specification to make it unambigous.
Did you think about filtering the environment variables passed to the
alert scripts? NOTIFY_SOCKET probably shouldn't be present, and PATH
probably shouldn't contain sbin directories; I guess all these are
inherited from systemd in my case.
I was also hit again by the "strange umask" problem here. It's set to
0026, which tends to get where nobody expects it (see for example
https://bugs.launchpad.net/fuel/+bug/1397284,
http://clusterlabs.org/pipermail/users/2015-June/000682.html, or
http://bugs.clusterlabs.org/show_bug.cgi?id=5268). In practice, alert
scripts won't often create local files, but it's a pity we have to fight
fallout from the logfile creation again.
(BTW I'd prefer to run the alert scripts as a different user than the
various Pacemaker components, but that would lead too far now.)
The SNMP agent seems to have a problem with hrSystemDate, which should
be an OCTETSTR with strict format, not some plain textual timestamp.
But I haven't really looked into this yet.
--
Regards,
Feri
More information about the Users
mailing list