739080 - redis stopped using /etc/redis/default.conf

Status: RESOLVED WONTFIX

(Infrastructure & Operations Graveyard :: CIDuty, task)

Description

[Chris AtLee [:catlee]]

at some point recently (around midnight today) redis on redis01 stopped using the configuration file located at /etc/redis/default.conf which is managed by puppet, and started using /etc/redis.conf. The critical difference between these two is that /etc/redis.conf only listens to localhost for connections, whereas /etc/redis/default.conf listens to all interfaces.

I've modified /etc/redis.conf to comment out the 'bind 127.0.0.1' line for now.

Comment 1

[Chris AtLee [:catlee]]

from /var/log/messages.1:

Mar 25 00:41:17 redis01 yum: Updated: redis-2.4.8-2.el5.x86_64

this is why ensure=>latest is EVIL

http://hg.mozilla.org/build/puppet-manifests/file/4a9544063c2b/modules/redis/manifests/install.pp#l4

Comment 2

[Phil Ringnalda (:philor)]

For bean-counting purposes, the combination of this and bug 739032 had the trees closed from 10:15 to 12:30 (though it should have been more like 00:01 to 12:30).

Comment 3

[John Hopkins (:jhopkins)]

I gather what we need to do is update puppet to use the new /etc/redis.conf location and stop using ensure=>latest on the redis package.

Comment 4

[Justin Wood (:Callek)]

(In reply to John Hopkins (:jhopkins) from comment #3)
> I gather what we need to do is update puppet to use the new /etc/redis.conf
> location and stop using ensure=>latest on the redis package.

Currently the redis server is not managed by puppet, it was oldPuppet.

that said, :catlee, is there anything more todo here other than recreating a Cent6-based redis puppetized system. If not then this bug should be resolved somehow

Comment 5

[Chris AtLee [:catlee]]

nope, all we need is a new host and really simple puppet module for it.

Comment 6

[Nick Thomas [:nthomas] (UTC+12)]

See also bug 863268.

Comment 7

[Chris Cooper [:coop] (he/him)]

I think this is WONTFIX if we're moving away from redis.