993496 - point [xxx] repository at new bm-remote webserver cluster to ensure parity in talos numbers

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

we currently use bm-remote for our talos on android to serve webpages.  We have a new updated fancier web multi headed cluster thingamajiggy.  Lets edit our configs in buildbot/mozharness to use the new server cluster for a given branch.  

Once we verify stability and parity with the old one, we can covert all mobile talos traffic to use this new web server.

Comment 1

[Kim Moir [:kmoir] ET]

So is there a specific branch you'd like to use for this testing?  cedar? ash?  What's the actual hostname of the new webserver?

Comment 2

[Joel Maher ( :jmaher ) (UTC -8)]

any branch is fine- I have been working on cedar lately.  

maybe dustin knows the new webserver

Comment 3

[Dustin J. Mitchell [:dustin] (he/him)]

It's http://talos-remote.pvt.build.mozilla.org

Comment 4

[Joel Maher ( :jmaher ) (UTC -8)]

looks like Dustin didn't clear the need-info.

:kmoir, is there anything else you need before getting this ready to switch?

Comment 5

[Kim Moir [:kmoir] ET]

No I'm writing/testing patches :-)

Comment 6

[Kim Moir [:kmoir] ET]

Attachment: mozharness patches

for android talos.  Yes, there are other links in the other android configs for the hostutils url but I think this is a good test, we can update the others later if the tests on cedar go well.

I've tested these patches on on cedar on my dev master and ensured that the correct values for the webserver are run on cedar, and then a sendchange against m-c includes the old value

Comment 7

[Kim Moir [:kmoir] ET]

Attachment: bug993496bb.patch

builbot config changes.  Not pretty but we should be able to just change the default value once the tests on cedar are complete.

Comment 8

[Armen [:armenzg]]

Comment on attachment 8405607 [details] [diff] [review]
bug993496bb.patch

You gotta love our configs :)

Comment 9

[Armen [:armenzg]]

Comment on attachment 8405605 [details] [diff] [review]
mozharness patches

Will the follow up patch will undo the "cedar" exception and remove the "#kim" comment?

Thanks for taking care of this bug!

Comment 10

[Kim Moir [:kmoir] ET]

Comment on attachment 8405605 [details] [diff] [review]
mozharness patches

yes the next patch will remove the cedar loop, and I removed the #kim comment this was there by mistake

Comment 11

[Kim Moir [:kmoir] ET]

in production with today's reconfig

Comment 12

[Kim Moir [:kmoir] ET]

verified that this is working on cedar, let me know when you want me to change this for real for all branches

Comment 13

[Kim Moir [:kmoir] ET]

Attachment: buildbot configs for when this moves to production

Comment 14

[Kim Moir [:kmoir] ET]

Attachment: mozharness patches for when this moves to production

Comment 15

[Kim Moir [:kmoir] ET]

Attachment: tools patches for when this moves to production

Comment 16

[Joel Maher ( :jmaher ) (UTC -8)]

Everything is good except for tp4m.  is there a possibility that we do not have the pageset for tp4m installed?

Also will there be docs and a process for updating the webserver?  We need to do that occasionally.  Here are the current docs:
https://wiki.mozilla.org/ReleaseEngineering:Buildduty:Other_Duties#Update_mobile_talos_webhosts

here is the sources for tp4m:
http://people.mozilla.org/~jmaher/mobile_tp4.zip

we expect to fine a url like this:
http://localhost/page_load_test/mobile_tp4/news.google.com/news.google.com/index.html

^ s/localhost/hostname/

Comment 17

[Amy Rich [:arr] [:arich]]

I took a look, and the symlinks were just bad for mobile_tp4 and tp4.  I've fixed those up.

https://wiki.mozilla.org/ReleaseEngineering:Buildduty:Other_Duties#Update_mobile_talos_webhosts is how to update things for the new hosts.

Comment 18

[Joel Maher ( :jmaher ) (UTC -8)]

thanks :arr!  I have verified the tp4m numbers are good for 2.2 and 4.0.

:kmoir, go ahead and transition over to this new webserver!

Comment 19

[Armen [:armenzg]]

Comment on attachment 8408344 [details] [diff] [review]
buildbot configs for when this moves to production

Review of attachment 8408344 [details] [diff] [review]:
-----------------------------------------------------------------

Great! Thanks Kim!

Comment 20

[Joel Maher ( :jmaher ) (UTC -8)]

next steps?  possibly just a reconfig?

Comment 21

[Kim Moir [:kmoir] ET]

yes just waiting for a reconfig

Comment 22

[Kim Moir [:kmoir] ET]

verified on tbpl

Comment 23

[Dustin J. Mitchell [:dustin] (he/him)]

Can you open up a bug to decom the old talos web hosts (the minis + netscaler)?  Or am I overly optimistic that they're now unused?

Comment 24

[Dustin J. Mitchell [:dustin] (he/him)]

I should read all my bugspam before replying.  That's bug 940551.  Nothing to do here :)

Comment 25

[Kim Moir [:kmoir] ET]

Aki: Can you update your mh repo with the most recent changes?  Talos tests on ash are failing because bm-remote has been decommissioned and ash uses your user repo which does not include this changes in this bug.

Comment 26

[Aki Sasaki (not active)]

  501  cd /src/clean/users/asasaki_mozilla.com/ash-mozharness/
  502  hg pull -u
  503  hg pull http://hg.mozilla.org/build/mozharness
  504  hg merge
  505  hg heads | more
  506  hg merge -r default

(in another terminal)
  502  vi mozharness/mozilla/testing/testbase.py
  503  hg resolve --mark mozharness/mozilla/testing/testbase.py
  504  hg commit
  505  hg push