UC upgrade: nova_db_sync hangs if nova cell DB connection string is no longer valid

[Jiří Stránský]

Right now we don't enable TLS in undercloud upgrade yet due to issues with Nova, and point all intra-undercloud traffic at 192.168.24.1. But even then another issue with Nova can appear - the DB connection strings for Nova Cells are in DB, not in config file. So nova_db_sync container will try connecting to 192.168.24.3 and hang (retry endlessly).

bash-4.4$ nova-manage cell_v2 list_cells
+---------+--------------------------------------+-------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------+----------+
| Name | UUID | Transport URL | Database Connection | Disabled |
+---------+--------------------------------------+-------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------+----------+
| cell0 | 00000000-0000-0000-0000-000000000000 | none:/ | mysql+pymysql://nova:****@192.168.24.3/nova_cell0?read_default_file=/etc/my.cnf.d/tripleo.cnf&read_default_group=tripleo | False |
| default | d1a18664-dde3-46b4-ae00-77047aa26dad | rabbit://guest:****@undercloud-0.ctlplane.localdomain:5672/?ssl=0 | mysql+pymysql://nova:****@192.168.24.1/nova?read_default_file=/etc/my.cnf.d/tripleo.cnf&read_default_group=tripleo | False |
+---------+--------------------------------------+-------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------+----------+

Workaround is likely updating the nova cell DB connection string before starting to upgrade undercloud (didn't test it yet).

[Jiří Stránský (jistr)]

I believe this no longer happens, will reopen if it does.