redis-server stops around 24 hours

[bjjay]

Do you find this problem? I don't know why, redis-server process isn't existed, but the files redis.pid and redis.sock are still existed.

[razorinc]

I think is something related with memory. I'm currently investigating this problem with guys from openshift on freenode...

[snitramodranoel]

Are there any updates on this issue? I'm very interested in deploying Redis with my application.

[bjjay]

No update on this issue, I guess.

[razorinc]

@snitramodranoel @bjjay I forgot to update the bug, i solved with a trivial cron job that check if the redis-server is running or not. But anyway, I think the openshift team is already working on a cartridge for redis :)

[bjjay]

Greet!

[snitramodranoel]

Thanks for the tip. Anyways, there's a poll about a redis cartridge on the OpenShift community web site. Maybe you guys already know about it. Let's spread the word and ask for votes ;)

[snitramodranoel]

@razorinc I suppose the issue is due to how OpenShift is implemented with SELinux. I'm not an OpenShift developer, but I bet this will only be solved for real when a redis cartridge is out. That's my two cents.

[razorinc]

@snitramodranoel Leonardo, I don't think is SELinux itself, but for sure, when the cartdrige will come out, would be way more easy to use it :) //cc @marekjelen

[rw950431]

Still no cartridge it seems. To add more details to @razorinc's answer.

Add cron cartridge to your app using procedure at https://openshift.redhat.com/community/blogs/getting-started-with-cron-jobs-on-openshift

script file in .openshift/cron/hourly/redis-check.sh containing
#!/bin/sh
kill -0 cat ${OPENSHIFT_RUN_DIR}/redis.pid || ${OPENSHIFT_DATA_DIR}/redis/bin/redis-server ${OPENSHIFT_DATA_DIR}/redis/bin/redis.conf

Then
$ git add .openshift/cron/hourly/redis-check.sh
$ git commit -a -m "add redis-server check cron job"
$ git push origin

[rw950431]

Corrected about script since Openshift fiddled with their enviroment variables see here