MariaDb Issues - anyone else?

Configuration
1

So I have been having some issues with my home assistant which I have eventually realised is because of the hass.io mariadb addon.

The first thing I have been seeing is that I have a package with themes.yaml in it with around 30+ themes in it. WHen 0.99.0 was released, I edited this file to add some extra variables so that the unused devices screen would display correctly. As soon as I did that, I started getting a database error when I restarted HA with it having a problem writing the theme to the database. I found that it was ok depending on how many of the themes I edited with the variable. Odd.

Yesterday, I decided to use a themes directory and split up the package so every theme was in it’s own package and I again edited all those themes for the new variables. The dB error returned when I did that.

In addition to this, regularly during the day, my home assistant was becoming unresponsive for a minute or so at a time. It always came back but it was happening all the time. Very annoying.

Lastly, a lot of my history graphs were flatlining after a day or so. I have recorder setup to not record 90% of entities etc so this was also odd.

Anyway, seeing the database errors yesterday I decided to switch back to SQLite dB - the HA default. As soon as I commented that out, all the above errors went away.

Has anyone else here had problems with the mariadb addon? I guess I can re-enable it and get the logged errors (and log an issue on github) but curious if it’s just me or not?

2

Not using it myself - but wondering if it a resource issue?

Disk space / quota?
memory for mariadb?

3

It’s running on a NUC with a 240gb SSD (40% free space) 8Gb RAM. It’s a hass.io addon so I can’t screw with any settings anyway.

4

That’s worrying. I haven’t noticed any DB errors (similar system specs to yours David) but will keep an eye on it.

5

I also should have mentioned I was getting a HA restart time of 5 minutes on a NUC whereas it used to be less than a minute.

In my mind linked it to a HACS update but was never really sure. Today I updated the samba addon and restarted HA and bugger me if it wasn’t back down to a minute. I think there was a HACS update yesterday as well… in any case, like lightning now again on my NUC. I suppose it’s also possible it was a docker update I installed today but can’t be certain about anything.

in any case thrilled it’s back to a minute or less. Anyone else seen this?

So maybe not a mariadb issue afterall. Hmm.