D4Science Infrastructure

This ticket track the "Rehearsal" of Production Jackrabbit Migration from Derby to PostgreSQL related to the steps listed in #8562

the migration process is now working with no exception but we have some issues in starting up the repo after the migration is completed...

the issues in starting up the repo after the migration is completed have been fixed by running a consistency check

Just to be sure that everything is provisioned, both machines workspace-repository-prod1.d4science.org and workspace-repository-prod.d4science.org are going to be recreated with ansible from scratch. Look at the sub-ticket with the details of the configurations. @tommaso.piccioli@isti.cnr.it already created a snapshot this morning which will be used for the latest (we hope) test, before the real migration scheduled for next week.

migration test has restarted at 15:50 23/06/17

migration ended at 22:39 23/06/17
consistency check/fix start running at 8:37 24/06/17

jackrabbit restarted at 15:50 (24/06/17) after consistency check and fix (it seems ok)

the overall time to complete the procedure is much higher than before, about 7 h for the migration and about 6h for the consistency check? I'm guessing the DB isn't local right? @costantino.perciante@isti.cnr.it

it has been still used a local db. Perhaps both VMs have less RAM/CPU than before @andrea.dellamico@isti.cnr.it ?

Costantino Perciante wrote:

it has been still used a local db. Perhaps both VMs have less RAM/CPU than before @andrea.dellamico@isti.cnr.it ?

Nope. The VM is the same, I only (re)installed smartgears and the home library webapp.

Massimiliano Assante wrote:

the overall time to complete the procedure is much higher than before, about 7 h for the migration and about 6h for the consistency check? I'm guessing the DB isn't local right?

The DB was local. It will moved to the remote server after the real migration.

I still don't know why a remote DB should be slower than a local one, btw.