Server down?

[bensenseii]

Actually, I think it should work since cgoban shows those lines "Due to the several hours long server outage on Dec 18th, the KGS Plus lectures have been cancelled on that day. We'll reschedule some event in January. We are sorry for the trouble!". However, it still doesn't work. That's quite boring and ununderstandable o.o

[glue]

KGS is back up !

[stalkor]

i hope you used the right glue this time:)

[emerus]

Did server just go down? If so, bad timing - right in the middle of an intense ASR game.

[ZeroKun]

Did server just go down? If so, bad timing - right in the middle of an intense ASR game.

Yes, unfortunately it is down again, wms is working on it.

[wms]

The server is indeed down again. The tech at the ISP looked and said that the memory was bad - this was probably the cause of the previous failure also. He replaced the memory and turned it back on again.

Right now the KGS database is being checked for errors (that is necessary whenever the OS dies intstead of getting a proper shutdown). I think it is most of the way done with the check, it should be up soon. Unless the check finds unfixable errors - then it could take me a long time (hours) to revert to backup. Luckily I have a backup from a few milliseconds before the crash...mysql replication is awesome for that.

[oren]

The server is indeed down again. The tech at the ISP looked and said that the memory was bad - this was probably the cause of the previous failure also. He replaced the memory and turned it back on again.

Right now the KGS database is being checked for errors (that is necessary whenever the OS dies intstead of getting a proper shutdown). I think it is most of the way done with the check, it should be up soon. Unless the check finds unfixable errors - then it could take me a long time (hours) to revert to backup. Luckily I have a backup from a few milliseconds before the crash...mysql replication is awesome for that.

Thanks for letting us all know what happened. It's interesting to see the problems you run into running the server.

[kirkmc]

100 bucks of hardware brings down a server... Isn't that something?

[wms]

Still checking the database...the biggest table (the one that maps players into games) has taken over an hour so far and still isn't done. But at least no errors reported yet in the database logs.

[jts]

Why would a backup that takes a few milliseconds to read off the server take a few hours to write back onto the server? (Not impatient, just curious.)

[wms]

Well, in this case it wouldn't be hours, but the difference is because the backup is done instantly - each database command that makes a change is sent to the backup server as soon as it runs. But then when you need to replace the master database, you need to copy the whole multi-gigabyte database from the backup to the master. I have two backups; one on the web server, which would copy over fairly quickly. Another at my home that would take hours to upload. I was thinking of my home backup, but since it was only the master and not the web server that had trouble, I could have gotten a copy of the database from there.

In case case, the server is up. Although the database was fixable, it was screwed up enough so that the web server lost its sync with the main database, so there is no rank system (it runs on the web server) and the web archives will not update. Tonight when I get home I'll pull them back into sync again and then things will be back to normal.

[daal]

Luckily I have a backup from a few milliseconds before the crash...mysql replication is awesome for that.

That is awesome!

[cdybeijing]

February 21 5:05 pm GMT +8 - server is refusing connections.

Is the server down or is this just me?

[Li Kao]

Doesn't work for me either.

[Jordus]

I get "Unable to Launch Application"...