[mythtv-users] mythconverg database server migration

Erik Karlin e_karlin at yahoo.com
Wed Mar 29 15:09:09 UTC 2017

I'm attempting to use the mythconverg_backup.pl / mythconverg_restore.pl
scripts to move my database from one very old slow server to another
old, but not as slow server. After a backup, my restore has been running
for about 6 days now, with the following output from show processlist:

| Id | User   | Host                    | db          | Command | Time | State  | Info
| 17 | mythtv |            myhost:37979 | mythconverg | Query   | 270851 | update | INSERT INTO `recordedseek` VALUES (1017,'2016-01-09 06:56:00',420,8972010,9),(1017,'2016-01-09 06:56 |

Now, checking my mythweb statistics, I see:
Total Running Time: 13 years 16 days 22 hrs 26 mins

so I know there is an awful lot of data to restore, but is there
anything I can do. I did a clean install of mysql 5.6 on an updated
debian jessie...no tuning or config changes other than allowing LAN

This was an empty database. I ran:
mythconverg_restore.pl --hostname dbhost --create_database --verbose

A few questions...
1. will this ever end?
2. can I kill this and restart, potentially exclude the big tables?
3. is there any tuning to be done?

Again, this was a fresh install of mysql 5.6. From a quick search, there
were some suggestions to turn off
autocommit/unique_checks/foreign_key_checks and something about
tuning key_buffer_size.

I assume if I can kill the restore, I can drop and recreate the database
without any problems, but is there a faster way? What if I don't load
the recordedseek table at all? I'd certainly prefer to not have to
interfere in a database load, but I need this to finish in some
definition of reasonable amount of time.

Any thoughts? The restore is still running if that matters.

Thanks in advance for any input

More information about the mythtv-users mailing list