[Raion]

I am working on a plan to migrate our forum to PostgreSQL. I've talked about it in the past, and I've found a few utilities to make this possible. 

Doing so will allow us to ditch MariaDB eventually and move to a proven database. 

It's not 100% yet, so I don't have a date/time window. We will have a Telegram channel, however, for users to get live updates on the site (and in the future, as well). 

I will notify everyone with a week's notice, hopefully, once I finish testing the dev site and extensions.

[Raion]

There are a few limitations that postgres will likely temporarily introduce but the main concerns are:

1. Maintain or improve our current performance characteristics.

2. Improve or maintain our search performance. I know the search is kinda meh as is. I'm looking at Sphinx.

3. Making sure all plugins work. I'm currently working on this in a test site.

Currently I can make a basic migration "work", but a few plugins break in the process. I will continue to look into fixing them. Once I've confirmed everything is A-OK I'll announce when.

[Raion]

Just to update everybody briefly, I've managed to fix a couple of plugins. However there are still some problems that remain that I'm going to have to work out over the next couple of weeks. Combined with actually juggling my personal life and going back to development of IRIX I don't have a timeline. Just know that our infrastructure is stable. I'm just working on this in my spare time as part of a personal journey to fix this stuff.

[Raion]

Postgres migration test hit a bunch of snags. Because I'm anal retentive, I'm already documenting a protocol of reversal in case stuff breaks. So this will take time.

[Raion]

A large setback has occurred -- the migration utils MyBB offers (Merge) won't preserve forum IDs/user IDs. Because our SEO relies on numbers, this is indefinitely postponed.

One idea is to setup a permalink style using forum topic names, and then let google slowly reindex for 6-8 months, but I'm not sure I wanna do that. We did take a minor hit in forcing moves off the old domains, not sure I'm prepared to undertake that.

[vishnu]

A large setback has occurred -- the migration utils MyBB offers (Merge) won't preserve forum IDs/user IDs. Because our SEO relies on numbers, this is indefinitely postponed.

One idea is to setup a permalink style using forum topic names, and then let google slowly reindex for 6-8 months, but I'm not sure I wanna do that. We did take a minor hit in forcing moves off the old domains, not sure I'm prepared to undertake that.

I've never trusted MariaDB either, not for any particular reason other than probably paranoia, but I'm still using a mysql on my database, which is 228,402,745 bytes, I mean, that's not really huge but still, it performs really well and I'm perfectly happy with it. But hey, that's just me, and YMMV! 😀

[Raion]

Oracle MySQL is essentially identical in a lot of respects to MariaDB. Biggest issue with Oracle MySQL is recent releases have broken features.

[vishnu]

Oracle MySQL is essentially identical in a lot of respects to MariaDB. Biggest issue with Oracle MySQL is recent releases have broken features.

Yes, I agree, the version of mysql that I use is from Slackware 14.2, which at this point is years old and has security vulnerabilities, but my database is on the darkweb and only selected viewers have the URL, and I watch the weblog to see if anyone I don't know has found access to it, and so far no one has. But either way I have the database backed up on a different computer, so if someone does manage to hack it I can restore it, and then worry about security updates..

[Raion]

Compared to my more youthful times, I certainly respect sticking with what you know and sticking with older versions of programs. I certainly have some systems that are running for five plus years at this point without a reinstall or major upgrade.

I do have to constantly react to the changes in the climates of things however. For long-term extensibility purposes it is going to be necessary to try to unify as much of our stack as possible. It may take a while for the postgres migration at this point as I am quickly losing my patience with the dev version of the site. If it has to wait it has to wait.

[vishnu]

Compared to my more youthful times, I certainly respect sticking with what you know and sticking with older versions of programs. I certainly have some systems that are running for five plus years at this point without a reinstall or major upgrade.

I do have to constantly react to the changes in the climates of things however. For long-term extensibility purposes it is going to be necessary to try to unify as much of our stack as possible. It may take a while for the postgres migration at this point as I am quickly losing my patience with the dev version of the site. If it has to wait it has to wait.

I don't really know that much about object-relational databases, and in general I'm not all that in favor of object oriented programming. I (we) do use C++ for all our programming at my place of employ, but mostly only because of the wonders of the standard template library. We do have some objects but mostly we stick to generic programming in the manner typified by Alexander Stepanov. Like, if you look at my ridiculous blackjack simulation that I posted here way back when, you can see that all the heavy lifting in the code is done by the STL, and there are zero c++ classes. Also, while I'm on the topic, I've since fixed the bugs that were found by jwhat and robespierre thanks for pointing those out guys!

I recently finished a pretty big run of the code and posted about it on Stanford Wong's blackjack site, for anyone who's bored enough to have a look:

https://bj21.com/boards/free/sub_boards/...e=1#unread