cinder-manage db sync fails due to row size too large

[Jared Baker]

OS: Ubuntu 20.04
DB: MariaDB 10.3
Release: Xena

Running the cinder-manage db sync to upgrade to Xena from Wallaby resulted in error:

2022-04-11 15:42:20.555 8601 CRITICAL cinder [-] Unhandled error: sqlalchemy.exc.OperationalError: (pymysql.err.OperationalError) (1118, 'Row size too large. The maximum row size for the used table type, not counting BLOBs, is 8126. This includes storage overhead, check the manual. You have to change some columns to TEXT or BLOBs')

The fix was to manually change row_format from 'Compact' to 'Dynamic' by running: ALTER TABLE volumes ROW_FORMAT=DYNAMIC;

We likely hit this issue because we are a long running cluster from Kilo era. I believe at some point the cinder-manage db sync should have changed this ROW_FORMAT to 'Dynamic' for us.

[Sofia Enriquez (lsofia-enriquez)]

This bug was discussed at the bug meeting: https://meetings.opendev.org/meetings/cinder_bs/2022/cinder_bs.2022-04-13-15.02.log.html#l-20

This is NOT A BUG. However, Cinder should improve the documentation and add a statement regarding the versions we test with/support. In particular, the newer mysql versions don't hit this.