Page 1 of 1

Updating ViciDial

PostPosted: Tue Mar 20, 2012 4:31 pm
by spacejanitor
Hello there,

I'm currently running a vicidial cluster

(VERSION: 2.4-357a
BUILD: 120125-2107)

and would like to know how people keep theirs updated? Of course for a production environment you would want to be careful with how/when your server(s) are updated... but is anybody running an auto-update from SVN, or is this even possible?

If you do choose to update... how do you do it, and what steps do you take to ensure the latest version is completely stable before deploying?

PostPosted: Sat Mar 31, 2012 10:32 pm
by williamconley
Upgrade instructions are in the Vicidial Wiki.

The method is simple enough: download the latest SVN or branch. Use the freshly downloaded "upgradeXXXX.sql" file to update your db. Then install Vicidial on top of the existing Vicidial (with install.pl). *poof*.

Suggestions:

We like to delete all the .sql lines that have "already executed" based on existing DB schema (which is updated for each release ... so you check your system settings, find your db schema, and delete all the already run lines).

We also recommend (highly) opting-in for the new sample files in case The Vicidial Group has updated any of them.

Stability:

We have not had ANY stability issues with the 2.4 trunk since shortly after we began updating to it. However, we did wait for a few months before we began (to let it mature ...). At present, 2.4 seems very stable at any install moment. No guarantees, however, which is why 7 days of backups is a good thing.

To date, we've never encountered any issues, but if we did ... it would likely be first thing at the beginning of shift and we'd restore the prior DB and reinstall the prior version of Vicidial and go back to work. Just be sure you're paying attention at the beginning of shift.

We do test simple autodial and upload, modify type functions after upgrade. But let's be real: You can't test EVERYTHING. LOL

PostPosted: Mon Apr 02, 2012 4:06 pm
by spacejanitor
Thanks for your feedback williamconley.

How about if you are running a cluster as opposed to a single unit? Would you do this same procedure for each server (DB, telephony, apache)? If so, in what order?