I have come across a strange behavior of the script
/etc/ha.d/resource.d/SendArp , packaged as a part of heartbeat
(heartbeat.x86_64 0:3.0.4-2.el6) on CentOS release 6.5, but the same
problem occurs at least on Debian in heartbeat-3.0.5 and presumably on
other distributions, as it is present in the 3.0.5 source tarball in
resources/heartbeat/SendArp.in).
If the OCF_TRACE_RA environment variable is set to 1 (OCF script tracing
turned on), the script works. The problem is that if OCF_TRACE_RA is not
set or set to 0 (the normal production setting), the script fails.
The script sources /etc/ha.d/resource.d/hto-mapfuncs which sources
/etc/ha.d/shellfuncs which in turn sources
/usr/lib/ocf/lib/heartbeat/ocf-shellfuncs. In the latter the last line is
ocf_is_true "$OCF_TRACE_RA" && ocf_start_trace
As the top level script (/etc/ha.d/resource.d/SendArp) sets the "exit on
error" shell option (set -e) at the beginning, the execution is
terminated upon the above statement for OCF_TRACE_RA unset or set to false.
I assume the 'set -e' command has just been forgotten there from a
debugging session and should be removed. Is that correct, or am I
missing something?
Thank you,
Martin
_______________________________________________
Linux-HA mailing list
Linux-HA@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
/etc/ha.d/resource.d/SendArp , packaged as a part of heartbeat
(heartbeat.x86_64 0:3.0.4-2.el6) on CentOS release 6.5, but the same
problem occurs at least on Debian in heartbeat-3.0.5 and presumably on
other distributions, as it is present in the 3.0.5 source tarball in
resources/heartbeat/SendArp.in).
If the OCF_TRACE_RA environment variable is set to 1 (OCF script tracing
turned on), the script works. The problem is that if OCF_TRACE_RA is not
set or set to 0 (the normal production setting), the script fails.
The script sources /etc/ha.d/resource.d/hto-mapfuncs which sources
/etc/ha.d/shellfuncs which in turn sources
/usr/lib/ocf/lib/heartbeat/ocf-shellfuncs. In the latter the last line is
ocf_is_true "$OCF_TRACE_RA" && ocf_start_trace
As the top level script (/etc/ha.d/resource.d/SendArp) sets the "exit on
error" shell option (set -e) at the beginning, the execution is
terminated upon the above statement for OCF_TRACE_RA unset or set to false.
I assume the 'set -e' command has just been forgotten there from a
debugging session and should be removed. Is that correct, or am I
missing something?
Thank you,
Martin
_______________________________________________
Linux-HA mailing list
Linux-HA@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems