366674 - Improvements to nagios tinderbox watching script

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Nick Thomas [:nthomas] (UTC+12)]

Attachment: Don't clobber in the absence of new info

This bug is to document improvements to the check_tinderbox_status script. 

We already landed this patch to prevent the old state being overwritten when the quickparse.txt file can't be retrieved. So means that there isn't a mail saying all the machines have fallen off, then another adding them all again when quickparse can be pulled. r=rhelmer on the changes.

Comment 1

[Nick Thomas [:nthomas] (UTC+12)]

Attachment: More improvments

This patch:

* adds a time out variable, so that if quickparse.txt is slow to dl then curl can end before Nagios gives up. This prevents mails about "CHECK_NRPE: Socket timeout after 10 seconds.", instead exiting with status 0 (so no message) and a useful string. Chose 7 seconds for timeout (arbitrarily)
* If there are no changes, return some output so that the pages like the URL don't say "CHECK_NRPE: No output returned from daemon"
* update the Revision # & blame
* simplify the if block for added builds
* make the new if blocks in rev 1.3 use the same style as the existing file

Comment 2

[Robert Helmer [:rhelmer]]

Comment on attachment 251175 [details] [diff] [review]
More improvments

I like the timeout idea, nice cleanup on those "if" statements.