I had to do a server migration recently which got wildly out of control due to a non-responsive (and eventually a former) web host. I've got two sets of the databases with record splicing. This isn't world-ending though because I had to put out every possible fire there is a lot of corruption with most of the IPv6 binary data returning empty results. I still have to confirm collation though I hadn't changed it from the defaults before importing (I don't think Swedish/Latin 1, whatever 10.11 is set to but not 520_ci. The previous server everything was set to utf8mb4 and 520_ci for character sets and collations. Now many rows are returned empty for INET6_NTOA(ipv6) even though the binary data looks the same for the same id records between the old and new servers.
I want to keep the data as-is from the original server (I already have it exported and locally stored on my computer) and then merge the new stuff from the new server. What software, on Windows, would make this the least painful and time-consuming?
I'll not be able to reply for ~8-16 hours so I'll try to preempt any structure questions with id always for autoincrement and date/date * for date columns if that helps any.
Alternatively, if not necessary I have full root access to everything and I suspect there might be some collation/character set tweak to get things back to the way they were.
New server events:
- Upgraded MariaDB to 10.11.
- Imported the databases.
- Logged in to root at the terminal/CLI and ran the following queries:
- SET CHARACTER SET 'utf8';
- SET collation_connection = 'utf8mb4_uca1400_ai_ci';
- SET GLOBAL collation_connection = 'utf8mb4_uca1400_ai_ci';
- SET GLOBAL collation_database = 'utf8mb4_uca1400_ai_ci';
- SET GLOBAL collation_server = 'utf8mb4_uca1400_ai_ci';
Suggestions please?