upgrade process raises error when using multiple schemas in postgresql

[Kamyar]

We are using multiple trac environments each of which are placed in a postgresql database schema (Single database). The first time the upgrade took place successfully for one of the database schema. For the rest of environments it raises the error:

InternalError: current transaction is aborted, commands ignored until end of tra nsaction block

Even using --no-backup does not help.

[Russ Tyndall]

I will try to get you straightened out, but first I am going to need a bit more info.

I need versions for postgres, trac, the python postgres bridge you are using, and the version of this plugin that you are using. Also helpful would be to enable TracLogging and post a log of the upgrade failing.

We run mulitple tracs from a single PG database (each trac in its own schema) and this has always worked for us (just so you know that this is indeed a supported way to run these packages). That said, MANY errors in trac manifest as a failure to upgrade.

Once you have posted this information, hopefully we can get you running pretty quickly.

Russ

[Russ Tyndall]

Just FYI, I just experienced the bug that you wrote about. I version 1.2.3 definitely works for me (I had it fail on 1.2.2 then work after upgrading to 1.2.3). The only problem I had after updating was that the system had previously set the version in the system table for key custom_report_manager_version to 1 (even though it had failed the upgrade process). I solved this annoyance in 1.2.4 (so an upgrade should succeed now).

Hopefully you should have no more trouble with this. Sorry it took so long to address this, I had not experienced the error till #9629 got me to track this down and fix it. Feel free to reopen if this is not meeting your needs

Cheers and thanks for the bug report, Russ