Closed Bug 863836 Opened 11 years ago Closed 11 years ago

please create a new database for servo buildbot master

Categories

(Data & BI Services Team :: DB: MySQL, task)

x86_64
Linux
task
Not set
normal

Tracking

(Not tracked)

VERIFIED FIXED

People

(Reporter: bhearsum, Assigned: scabral)

References

Details

We're going to be creating a new buildbot master for the Servo project - one that won't be related to our existing buildbot db.

Let me know what other details you need.
Should this go on the buildbot cluster?  Other options are the generic and devtools clusters, and the not-yet-built releng cluster.  I'm not sure who I'm asking, here - Ben, Sheeri, Corey..
I'm trying to keep servo infrastructure as close to other RelEng infra to ease consolidation later so my intuition says "buildbot cluster", but I don't know all the implications of that.
Fixing component!
Assignee: server-ops-dcops → server-ops-database
Component: Server Operations: DCOps → Server Operations: Database
QA Contact: dmoore → cshields
Well, for now, how about I make it on the staging server? The implications of having servo on the same machine as buildbot is that buildbot failing is tree-closing.

What's the db name I should put in? something like servo_allizom_org?
Assignee: server-ops-database → scabral
This isn't a staging instance, so that doesn't seem like a good choice.  Ben, do the trees close if this fails?
(In reply to Dustin J. Mitchell [:dustin] from comment #5)
> This isn't a staging instance, so that doesn't seem like a good choice. 

+1

> Ben, do the trees close if this fails?

Servo's workflow is a little different, but the simple answer is "no". Servo can survive outages better than we can on the other Buildbot DB, but it's still going to be a production service...so long outages are bad, of course.
I know it's not a staging instance, but don't you want one, to test?
(In reply to Sheeri Cabral [:sheeri] from comment #7)
> I know it's not a staging instance, but don't you want one, to test?

Yeah - that's a good idea.
(In reply to Sheeri Cabral [:sheeri] from comment #4)
> Well, for now, how about I make it on the staging server? The implications
> of having servo on the same machine as buildbot is that buildbot failing is
> tree-closing.
> 
> What's the db name I should put in? something like servo_allizom_org?

Oh, and let's not use servo_allizom_org...how about servo_buildbot_staging?
How about buildbot_servo_staging that way we can do something like "show databases like "buildbot%" and see everything?
buildbot_servo_staging db and buildbot_servo user created, password given securely to Ben.
virtual IP for stage is:

writes: stage-rw-vip.db.scl3.mozilla.com      
reads: stage-ro-vip.db.scl3.mozilla.com
Can we get the prod the db created now?
OK, we can do either devtools or generic, pinging bhearsum in IRC about which he prefers (what might interfere or not play nicely with others).
I have created buildbot_servo database on the devtools cluster and gave the buildbot_servo user permissions on the database. Will send password securely.
VIPs:

 devtools-ro-vip.db.scl3.mozilla.com      	
 devtools-rw-vip.db.scl3.mozilla.com
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → FIXED
tested the rw credentials and they work fine. Thanks!
Status: RESOLVED → VERIFIED
Product: mozilla.org → Data & BI Services Team
You need to log in before you can comment on or make changes to this bug.