Closed Bug 1137912 Opened 9 years ago Closed 9 years ago

Please deploy pushgo 1.5rc4 to Loop-Push Stage

Categories

(Cloud Services :: Operations: Deployment Requests - DEPRECATED, task)

x86
macOS
task
Not set
normal

Tracking

(Not tracked)

VERIFIED FIXED

People

(Reporter: benbangert, Assigned: oremj)

References

Details

1.5rc4 has been verified under my load-testing and is ready to go to staging.
Assignee: nobody → oremj
Deployed to stage.
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
Richard, this is ready for QA verification before production deploy.
Flags: needinfo?(rpappalardo)
I should note, any ticket for production deploy opened by QA, will need some config changes first. Should these config changes be made first to stage? (I'm guessing probably so)
(In reply to Ben Bangert [:benbangert] from comment #3)
> I should note, any ticket for production deploy opened by QA, will need some
> config changes first. Should these config changes be made first to stage?
> (I'm guessing probably so)

Hi Ben, if they're changes that could impact deployment or application behavior then ideally yes. 
What config changes are these?
Flags: needinfo?(rpappalardo) → needinfo?(bbangert)
The bucket size needs to be increased to match the amount of servers in the cluster, and always route needs to be enabled.

:oremj, the changes to be made:

- Under [router]
  - bucket_size should match the size of the cluster (in production as well as staging)
- Under [endpoint]
  - always_router should be set to true

That should be it.
Flags: needinfo?(bbangert) → needinfo?(oremj)
Sorry, always_route, not always_router
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
I've deployed the config changes. In the future, it would be nice if bucket_size were dynamically configured via etcd, so if the cluster size changes, the bucket size also follows.
Flags: needinfo?(oremj)
Status: REOPENED → RESOLVED
Closed: 9 years ago9 years ago
Resolution: --- → FIXED
Ok, :rpapa, config changes are out. Ready for verification.
Flags: needinfo?(rpappalardo)
(In reply to Ben Bangert [:benbangert] from comment #3)
> I should note, any ticket for production deploy opened by QA, will need some
> config changes first. Should these config changes be made first to stage?
> (I'm guessing probably so)

Thanks, Ben.  Normally, I would create a prod ticket and assign to oremj once stage is verified.  It sounds like these config changes have already been applied on stage and we don't need any further changes there prior to deploying to prod (pls let me know if otherwise).  If so, will we be doing it that way going forward as well? Just want to make sure the deployment process is clear. 

Also, I assume 1.5rc4 will be deployed to prod as 'Loop-push 1.5'?


------------------
TEST SUMMARY
------------------

make test against tag - OK
(https://pastebin.mozilla.org/8824047)

loadtesting (per :benbangert) - OK

end-2-end loop call verification - OK
Placed several calls back and forth between Nightly and DevEd(Aurora).
exercising video/audio mute, join/leave call, window-sharing, etc. functions.

https://loop.stage.mozaws.net/push-server-config & stage load-balancing - OK


Once above is clarified, I'll verify stage.
Flags: needinfo?(bbangert)
Yep, that is correct, this will be Loop-push 1.5.
Flags: needinfo?(bbangert)
Thanks, Ben. I'll create a prod deployment ticket early next week right after loop-client is out the door.
Hey Jeremy, I'd like to coordinate a loop-push prod deployment tomorrow if you're available.  Would you be free in the late morning around 10:30?  If possible, I'd like to do it before noon in case we run into any issues.  If you're busy, we could plan on Thurs. morning instead.  Let me know what works for you.
Flags: needinfo?(oremj)
Yes, 10:30 works for me.
Flags: needinfo?(oremj)
Status: RESOLVED → VERIFIED
OK, great.  I'll ping you then.
Thanks!

ticket here:
Bug 1141860 - Please deploy pushgo 1.5 to Loop-Push Production
Blocks: 1141860
You need to log in before you can comment on or make changes to this bug.