963204 - slaveapi's reboot action shouldn't create a bug for a slave unless it was unable to reboot it

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[[e-mail] (:bhearsum)]

Just noticed this today when I saw https://bugzilla.mozilla.org/show_bug.cgi?id=963197 fly by.

Comment 1

[[e-mail] (:bhearsum)]

Attachment: fix slave bug creation logic

Comment 2

[[e-mail] (:bhearsum)]

Attachment: better patch

Per IRC, I've modified the patch to not load bug info twice. I switched up the state from REOPENED -> NEW because Bugzilla complains if you try to change a bug from NEW -> REOPENED (which is what happens when the IT bug is filed when the slave bug is already open).

Comment 3

[[e-mail] (:bhearsum)]

Attachment: the right patch

Comment 4

[[e-mail] (:bhearsum)]

Comment on attachment 8364595 [details] [diff] [review]
the right patch

Landed and bumped the version in Puppet.

Comment 5

[[e-mail] (:bhearsum)]

Landed and working.

Comment 6

[Justin Wood (:Callek)]

(In reply to Ben Hearsum [:bhearsum] from comment #2)
> Created attachment 8364576 [details] [diff] [review]
> better patch
> 
> Per IRC, I've modified the patch to not load bug info twice. I switched up
> the state from REOPENED -> NEW because Bugzilla complains if you try to
> change a bug from NEW -> REOPENED (which is what happens when the IT bug is
> filed when the slave bug is already open).

Unfortunately,  it seems this patch has stopped the reopening existing bugs.

Comment 7

[[e-mail] (:bhearsum)]

(In reply to Justin Wood (:Callek) from comment #6)
> (In reply to Ben Hearsum [:bhearsum] from comment #2)
> > Created attachment 8364576 [details] [diff] [review]
> > better patch
> > 
> > Per IRC, I've modified the patch to not load bug info twice. I switched up
> > the state from REOPENED -> NEW because Bugzilla complains if you try to
> > change a bug from NEW -> REOPENED (which is what happens when the IT bug is
> > filed when the slave bug is already open).
> 
> Unfortunately,  it seems this patch has stopped the reopening existing bugs.

Argh! It looks like RESOLVED -> NEW is also an invalid change:
 69947 2014-01-25 17:01:03,647 - ERROR - Something went wrong while processing!
 69948 Traceback (most recent call last):
 69949   File "/builds/slaveapi/prod/lib/python2.7/site-packages/slaveapi/processor.py", line 58, in _worker
 69950     res, msg = action(slave, *args, **kwargs)
 69951   File "/builds/slaveapi/prod/lib/python2.7/site-packages/slaveapi/actions/reboot.py", line 80, in reboot
 69952     slave.bug.add_comment(status_text, data={"status": "NEW"})
 69953   File "/builds/slaveapi/prod/lib/python2.7/site-packages/slaveapi/clients/bugzilla.py", line 16, in add_comment
 69954     return bugzilla_client.add_comment(self.id_, comment, data)
 69955   File "/builds/slaveapi/prod/lib/python2.7/site-packages/bzrest/client.py", line 66, in add_comment
 69956     return self.update_bug(id_, data)
 69957   File "/builds/slaveapi/prod/lib/python2.7/site-packages/bzrest/client.py", line 61, in update_bug
 69958     return self.request("PUT", "bug/%s" % id_, data)
 69959   File "/builds/slaveapi/prod/lib/python2.7/site-packages/bzrest/client.py", line 51, in request
 69960     raise BugzillaAPIError(resp["code"], resp["message"], resp)
 69961 BugzillaAPIError: (u'You are not allowed to change the bug status from RESOLVED to NEW.', {u'message': u'You are not allowed to change the bug status from RESOLVED to NEW.', u'code': 123, u'error': True})


There's 19 of them in the log, which I think accounts for all of them...

Comment 8

[[e-mail] (:bhearsum)]

Attachment: fix bug filing

This should fix the issue. I tested it in 4 scenarios against bugzilla-dev:
* No slave bug, no reboot bug: https://bugzilla-dev.allizom.org/show_bug.cgi?id=talos-r3-fed-096
* Open slave bug, no reboot bug: https://bugzilla-dev.allizom.org/show_bug.cgi?id=talos-r3-fed64-058
* Closed slave bug, no reboot bug: https://bugzilla-dev.allizom.org/show_bug.cgi?id=talos-r3-fed-098
* Open slave bug, open reboot bug: https://bugzilla-dev.allizom.org/show_bug.cgi?id=talos-r4-snow-023

The edits around 8:36/8:37 are me manually tweaking bugs to be in the state I needed them to be before testing the patch.

Comment 9

[[e-mail] (:bhearsum)]

This in production and slaverebooter has been turned back on. I'm doing a manual run of it now, to get caught up, then it will continue to run every 4 hours as before.