Closed Bug 947209 Opened 11 years ago Closed 11 years ago

Auto-create bugzilla tracking bugs for machines

Categories

(Release Engineering :: General, defect)

x86
macOS
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED DUPLICATE of bug 963204

People

(Reporter: pmoore, Unassigned)

Details

Currently we manually create tracking bugs such as https://bugzilla.mozilla.org/show_bug.cgi?id=bld-lion-r5-086

These could be manually created from a cron which inspects an agreed data source (such as inventory database) and finds new machines that currently do not have associated bugs.

Not sure if this also makes sense for cloud instances, or whether these are even in the inventory db (suspect not).

Also up for discussion if this should be limited to slaves, or also masters, or even all rel eng machines including those performing other services, or even all machines, not just the releng ones.

Anyway, at a minimum, would be nice to have releng slaves in there.

Pete
^^ These could be *automatically* created (not "manually) ... ^^^
^^^ "manually" ^^^
I think that bug 859403 might just be about reopening an existing tracking bug, triggered by a failure case. I was more thinking about a cron that is not checking failures, but instead queries bugzilla to determine a list of machines with tracking bugs, and queries inventory database (or slavealloc, or some other data source) and gets the list of build slaves (and maybe masters too) and then checks the two lists are equivalent. If a machine exists in inventory database, but does not have an associated (open or closed) tracking bug, it will create one - initially closed.

I believe bug 859403 is not checking an inventory database, it is looking for a list of failing machines, and then reopens (or even creates) the tracking bug - but it is useful to have the tracking bug before the machine starts failing, e.g. in case we want to loan it out. If we only create tracking bugs automatically when a machine fails, it means if we want to loan a machine that has never failed, we would still need to manually create the tracking bug...

Hope that clarifies. :)

Pete
(In reply to Pete Moore [:pete][:pmoore] from comment #4)
> I think that bug 859403 might just be about reopening an existing tracking
> bug, triggered by a failure case. I was more thinking about a cron that is
> not checking failures, but instead queries bugzilla to determine a list of
> machines with tracking bugs, and queries inventory database (or slavealloc,
> or some other data source) and gets the list of build slaves (and maybe
> masters too) and then checks the two lists are equivalent. If a machine
> exists in inventory database, but does not have an associated (open or
> closed) tracking bug, it will create one - initially closed.
> 
> I believe bug 859403 is not checking an inventory database, it is looking
> for a list of failing machines, and then reopens (or even creates) the
> tracking bug - but it is useful to have the tracking bug before the machine
> starts failing, e.g. in case we want to loan it out. If we only create
> tracking bugs automatically when a machine fails, it means if we want to
> loan a machine that has never failed, we would still need to manually create
> the tracking bug...

slaveapi is meant to do this, and already pulls in the various data sources (inventory, slavealloc) to make its bug-filing decisions.
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → DUPLICATE
Component: Tools → General
You need to log in before you can comment on or make changes to this bug.