598750 - Provide access to Socorro staging and production for devs

Status: RESOLVED FIXED

(mozilla.org Graveyard :: Server Operations, task)

Description

[Laura Thomson :laura]

As discussed at the 1.8 postmortem, devs should have access to staging.  The list is:
laura
lars
ozten
rsnyder
rhelmer
deinspanjer

Comment 1

[Laura Thomson :laura]

Just expanding this to cover read access to prod as well.

Jabba, can you give me an ETA on this please?

Comment 2

[Justin Dow [:jabba]]

I configured pm-collector-stage1 and dm-breakpad-stage01 to use ldap configs for ssh access. This means that anyone with a posixAccount attribute (test: can you log into khan?) should be able to log into these two boxes now.

Is sudo access required as well for the staging boxes?

I'm currently unsure about providing access to mrapp-stage02, which is where the webapp-php lives. If that is required, it might be better to move that to a different box.

Comment 3

[Laura Thomson :laura]

We should get access to the webapp.  I assume the hesitation is because other webapps are hosted on mrapp-stage02.  I'll defer to mrz here but I thought the general trend was to give access for webdevs to staging.

Comment 4

[Ryan Snyder [:ryansnyder] [:rsnyder] [:rysny]]

I was able to successfully login to both servers.  Thanks!

Comment 5

[Justin Dow [:jabba]]

(In reply to comment #3)
> We should get access to the webapp.  I assume the hesitation is because other
> webapps are hosted on mrapp-stage02.  I'll defer to mrz here but I thought the
> general trend was to give access for webdevs to staging.

That is correct. I think, though, that the general access to staging will be planned out and deployed at a later date, so I think for now I'll move the webapp to a different server.

Comment 6

[Justin Dow [:jabba]]

The webapp has been moved to pm-collector-stage1. So now dm-breakpad-stage01 and pm-collector-stage1 contain the entirety of stage.

Comment 7

[Justin Dow [:jabba]]

I've got users and ssh keys in puppet and pushed out to the collectors: still needed are the processor/monitor boxes and the middleware boxes. Also need ssh access opened up. Filed bug 605683 for that.

Comment 8

[Justin Dow [:jabba]]

We are going to run into the same issue that we ran into for stage here for the PHP webapp. It is on shared infrastructure. Not sure how to handle access there. 

mrz: how should this part be handled?

Comment 9

[matthew zeier [:mrz]]

Which infra is this?

Comment 10

[Justin Dow [:jabba]]

mrz: This resides on the generic cluster.

Comment 11

[Justin Dow [:jabba]]

Access has been granted to pm-app-collector01 - 06 ; cm-breakpad02, 03 and 04 and dm-bp-mware01. We are going to have to WONTFIX granting access to the servers that the reporter resides on.

Comment 12

[Laura Thomson :laura]

OK, what can we do to get better insight into the reporter?  Can we get logs mirrored to a box we do have access to?  (khan?)

Comment 13

[Justin Dow [:jabba]]

The kohana logs, or the apache logs?

Comment 14

[Laura Thomson :laura]

(In reply to comment #13)
> The kohana logs, or the apache logs?

Both?

Comment 15

[matthew zeier [:mrz]]

Don't the logs on mpt-vpn://var/log/clusterlogs get you want you want?