Since it was asked on another bug, these bots should be added to the public chromium-swarm instance.

Also, I would prefer if happening after Nov 4th as I'm OOO next week and whesse@ is traveling some of the days.

Updated title

Sorry for the delay.  How do you want this to be done?  Can I take them down, change their hostnames (they'll be -b9 hosts) and bring them up as swarming?  Also, since they're android hosts, will they be dockerized swarming?

Happy to work on this. Just want to verify post-holiday availability of dart folks before starting. Thanks.

-Sorry, I just returned from vacation. We can move forward with this now.

Can I take them down, change their hostnames (they'll be -b9 hosts) and bring them up as swarming?

-Yes, that is what we have done before.

Also, since they're android hosts, will they be dockerized swarming?

- I'd prefer we keep them similar to what they are currently to make it easier to make this work with our test framework as is and worry about dockerization later.

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chrome-golo/chrome-golo/+/608faa91ccbab3d0f65b2474dacd706f38019adc

commit 608faa91ccbab3d0f65b2474dacd706f38019adc
Author: John Weathersby <johnw@google.com>
Date: Thu Jan 10 00:13:07 2019

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chrome-golo/chrome-golo/+/1a7adb047a9395dc39f1fd9f6c305bcf3b7d2416

commit 1a7adb047a9395dc39f1fd9f6c305bcf3b7d2416
Author: John Weathersby <johnw@google.com>
Date: Thu Jan 10 00:44:59 2019

build{10,11,12}-b3 swarmed as build{20,21,22}-b9 (respectively)

Let us know if these require any further treatment, dockerization, etc. Thanks.

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/infradata/config/+/f63eea7968fcc1a421b3fe837ec4d0458d6cd291

commit f63eea7968fcc1a421b3fe837ec4d0458d6cd291
Author: Alexander Thomas <athom@google.com>
Date: Thu Jan 10 09:22:46 2019

This didn't quite work as I had expected because swarming quarantines the bots with multiple devices attached so they can't be used. So I think we'll have to dockerize after all. Something similar to how build19-b9 is setup should work for us as well.

We could just not use the android.py bots.cfg script, but that seems to do a lot of good stuff® to keep the devices healthy so I think we're better off using it.

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/infra/puppet/+/30bfe1cc3754aa20177f5246cea37fc587113ba9

commit 30bfe1cc3754aa20177f5246cea37fc587113ba9
Author: John Weathersby <johnw@google.com>
Date: Fri Jan 11 03:49:16 2019

These (dockerized?) android bots are now working (when they are online, only 1/2 are) as shard runners for our android testing, as we intended:

https://chromium-swarm.appspot.com/task?id=42568323f4f30510&refresh=10

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/infra/puppet/+/f13c1f5a0aacce2628b6ca987f4ef821ca41b20a

commit f13c1f5a0aacce2628b6ca987f4ef821ca41b20a
Author: John Weathersby <johnw@google.com>
Date: Fri Jan 11 19:16:45 2019

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/infra/puppet/+/62290519e80b400e6e6bf5c6f2e888a3150f36e8

commit 62290519e80b400e6e6bf5c6f2e888a3150f36e8
Author: John Weathersby <johnw@google.com>
Date: Fri Jan 11 19:33:50 2019

The dockerized bots work now, thanks for the quick fix!

Now, one last oddity is that two devices seem to be missing:
https://chromium-swarm.appspot.com/bot?id=build21-b9--device2
https://chromium-swarm.appspot.com/bot?id=build20-b9--device3

Hmm... Will have a look at those. Thanks.

build21-b9's devices all had older/flaky interconnects. swapped these all out.
In addition there was a device that was completely DOA. Flashed and provisioned a replacement, and it looks like these now have full complements. (full being 6 devices each in this case).

Let us know if everything looks ok.

Device stability seems acceptable. Closing this out.

Thanks!