Database error fixed on this forum today :)

[Tas_mania]

I noticed this forum was down for awhile today. These errors are quite common and are mostly mysql database errors. From memory they usually mean stopping and re-starting the database but they need expert attention. Congratulations to the admins at Gimp-Forum

[Ofnuts]

(10-22-2025, 08:17 AM)Tas_mania Wrote: I noticed this forum was down for awhile today. These errors are quite common and are mostly mysql database errors. From memory they usually mean stopping and re-starting the database but they need expert attention. Congratulations to the admins at Gimp-Forum

You don't even need to restart, just start a SQL session and issue REPAIR TABLE <the_bad_table>.

OTOH, what do you mean by "These errors are quite common"? Do you see them often here? I only recall fixing one such mishap about 6 weeks ago. Am I missing any?

[Tas_mania]

Quote:"These errors are quite common"? Do you see them often here?

Hi ofnuts, no I meant generally across the net. I once managed sites and they all use a DB of some sort. I don't know why they get corrupted. Maybe power supply drop-outs and a backup generator is slow to start. Tasmania gets really fierce winds at times that play havoc with TV and electrical systems. The electric company uses drones to check power lines and they find the problem faster

[Ofnuts]

(10-23-2025, 10:24 AM)Tas_mania Wrote:

Quote:"These errors are quite common"? Do you see them often here?

Hi ofnuts, no I meant generally across the net. I once managed sites and they all use a DB of some sort. I don't know why they get corrupted. Maybe power supply drop-outs and a backup generator is slow to start. Tasmania gets really fierce winds at times that play havoc with TV and electrical systems. The electric company uses drones to check power lines and they find the problem faster

Most likely a software error here. It's always the same table that gets corrupted, the one that keeps the sessions. Since we have about 10K users per day, that's the one that gets the most updates. A power outage that would be quite a rare occurrence in a data center located in the Netherlands, and would make the "virtual server" reboot, and I don't see any indication of this (according to uptime G-F.n has been up for several weeks).

[Tas_mania]

OK I get the idea. Could it be possible to reduce the amount of sessions data being collected? If there are 'peak' sessions times then this table could get overloaded. Also there could be zombie sessions of users that never logged out. Just my thoughts on it. It would be interesting to see just what is being included.

[Ofnuts]

(10-23-2025, 09:08 PM)Tas_mania Wrote: OK I get the idea. Could it be possible to reduce the amount of sessions data being collected? If there are 'peak' sessions times then this table could get overloaded. Also there could be zombie sessions of users that never logged out. Just my thoughts on it. It would be interesting to see just what is being included.

From the contents it automatically purges everything older than 24 hours. I don't think regular users are a problem here. AI bots are becoming a general nuisance.

[Tas_mania]

Quote:From the contents it automatically purges everything older than 24 hours. I don't think regular users are a problem here. AI bots are becoming a general nuisance.

I guess you could reduce the 24 hours down to a lower value. 24 hours is longer than people's short term memory   What was I doing 24 hours ago?

AI bots are really just free loaders IMHO. This is what AI thinks about AI bots;

   

They harvest images and also answers to questions. They can tell people how to block AI and possibly destroy AI which could be 'AI suicide'. Maybe AI 'voluntary euthanasia'?