1040109 - remove references to scl1 and 10.12 from puppetagain

Status: RESOLVED FIXED

(Infrastructure & Operations :: RelOps: Puppet, task)

Description

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

Now that we've moved all of the machines out of scl1, we should go through and remove all instances of scl1 and the 10.12 network from puppet.

Comment 1

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

Attachment: bug1040109.patch

Comment 2

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

Comment on attachment 8458823 [details] [diff] [review]
bug1040109.patch

Review of attachment 8458823 [details] [diff] [review]:
-----------------------------------------------------------------

What you include here looks accurate, but it also looks like you're missing these two files that also reference scl1 (the reason for the r-):

./modules/buildslave/files/runslave.py
./modules/buildslave/files/test_runslave.py


A couple other general comments (not review blocking, but "as long as you're cleaning things like redis up")... 

1) There are some people who should no longer have accounts because they've left the company.
2) There are a lot of node definitions where we could happily regex our way out of multiple lines of text.  e.g. b-linux64-{ix|hp} or lumping all of the os x try-level access talos nodes together by vlan (test.releng.scl3.mozilla.com).  Since they all include the same toplevel and set the same trust level, I'm not sure it's worth breaking them out by OS.

Comment 3

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

(In reply to Amy Rich [:arich] [:arr] from comment #2)
> Comment on attachment 8458823 [details] [diff] [review]
> bug1040109.patch
> 
> Review of attachment 8458823 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> What you include here looks accurate, but it also looks like you're missing
> these two files that also reference scl1 (the reason for the r-):
> 
> ./modules/buildslave/files/runslave.py
> ./modules/buildslave/files/test_runslave.py

That's mentioned in the commit message, actually.  That's dead code anyway (from before scl3) and needs to be removed entirely, but doing that removal runs the risk of causing buildbot startup to fail, so it's not a good change for this bug, which is otherwise very easy to test and very low-risk.  But, I can add it.

> A couple other general comments (not review blocking, but "as long as you're
> cleaning things like redis up")... 
> 
> 1) There are some people who should no longer have accounts because they've
> left the company.

Yeah, bug 926468 would help..

> 2) There are a lot of node definitions where we could happily regex our way
> out of multiple lines of text.  e.g. b-linux64-{ix|hp} or lumping all of the
> os x try-level access talos nodes together by vlan
> (test.releng.scl3.mozilla.com).  Since they all include the same toplevel
> and set the same trust level, I'm not sure it's worth breaking them out by
> OS.

This is a good cleanup bug - bug 1040926.

Comment 4

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

Attachment: bug1040109-p1.patch

Comment 5

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

Comment on attachment 8458892 [details] [diff] [review]
bug1040109-p1.patch

Review of attachment 8458892 [details] [diff] [review]:
-----------------------------------------------------------------

mea culpa for only reading the diff.  I took a look at the rest of the reunslave stuff.  Wow, what an ancient cesspool.  Is there a bug on file to amputate that entirely?