1336302 - decom anemometer: IT puppet webapp::anemometer and bits on scalebase1.db.scl3

Status: RESOLVED FIXED

(Infrastructure & Operations Graveyard :: WebOps: Other, task)

Description

[:Atoll]

Bug 862898 implemented Anemometer on the scalebase cluster in PHX1. Is this software still in use? If so, we need to fix this URL

modules/webapp/manifests/anemometer.pp:#       "hg clone http://hg.mozilla.org/qa/testcase-data/ .":

To clone from https://hg instead. I notice that it's commented out, so maybe it can just be removed instead.

(If anemometer is not in use any longer, then this should become a 'decom anemometer' bug.)

Comment 1

[:Atoll]

Clearing all secgroups, because this doesn't need to be secret.

Comment 2

[Matt Pressman [:mpressman]]

I'm not aware of our current use of this.

Comment 3

[:Atoll]

This is deployed as a website to scalebase1.db.scl3. If the weblogs on that host confirms it's unused, you're a winner.

Comment 4

[Ryan Watson [:w0ts0n]]

Nothing in the weblogs other that 127.0.0.1/server-status for a long time.

Doesn't look like it's used, I suggest stopping apache or the server and seeing if anyone yells loud enough, otherwise I can't see any usage.

Comment 5

[Ryan Watson [:w0ts0n]]

I didn't get time to fully investigate, but the parts I did investigate led me to the conclusion that nobody cares about these machines. I believe they are good to decom.

Comment 6

[:Atoll]

:mpressman, can scalebaseX.db.scl3 be decom'd as a group? If so, that simplifies this dramatically.

Comment 7

[Matt Pressman [:mpressman]]

Yep, these can be gone!

Comment 8

[chris valaas [:cvalaas]]

mysqld is now shutdown on scalebase[12].db.(scl3|phx1).mozilla.com

waiting on webops to weigh in on whether they need the newrelic db monitors which run on those hosts.

Comment 9

[:Atoll]

If the DB servers can be shutdown, it seems extremely unlikely that WebOps would need to run New Relic on the now-shutdown servers.

Shyam, rubber-stamp "permission to decom scalebaseX.db.scl3|phx1 *including* New Relic" requested?

Comment 10

[chris valaas [:cvalaas]]

To clarify: the newrelic monitors are monitoring a bunch of *other* databases (list sent to Shyam under separate cover), not the databases on the scalebase hosts themselves.

Comment 11

[:Atoll]

That makes more sense - thanks! Clearing needinfo, triage queue is enough.

Comment 12

[Eric Ziegenhorn :ericz]

Closing invalid as Scalebase is now shut down.  Thanks everyone.

Comment 13

[Eric Ziegenhorn :ericz]

Reopening for the removal of the puppet module, overlooked the forest for the trees there.

Comment 14

[Eric Ziegenhorn :ericz]

Removed anemometer-related puppet files in git commit 85d8f66f0805a262a534b4c2168b4fb8ec7bc79c.  There are still scalebase-related files in puppet/manifests/nodes/db_mysql/newrelic.pp which should be removed when we do the full decom, see bug 1442257.