Today's server problems

[exxos]

I’d like to take this opportunity to once again say “it wasn’t me.”

Maybe your hammer is wireless..

[exxos]

Thanks for all the hard work you do keeping this site going dude, it really is appreciated, and Merry Chrismas fella

Thanks,

[PhilC]

I’d like to take this opportunity to once again say “it wasn’t me.”

Maybe your hammer is wireless..

Well it certainly can be but unlike Mjornir, it doesn’t come back when called.

[exxos]

So it seems we are not out of the woods yet

Because I am running the wiki on exxoshost also as a test setup just to make sure things are fine.. Some pages are timing out.. I tried to delete the database in phpmyadmin and it timed out also.

So I am trying to DROP the wiki database from the command line where it has been stuck for several minutes now normally completes within just a couple of seconds.

So I fear something is wrong with the server, maybe it as a bad sector on the disk or something

I have opened up a ticket with ram node about this because something is clearly going on...

But also I'm not at home from tomorrow and I don't imagine there's going to be any support over the next few days. So if the server goes down during that time there , I'm afraid I won't be able to do anything about it

EDIT: it finally dropped the database:

Code: Select all

Database changed
mysql> DROP DATABASE wiki;
Query OK, 0 rows affected (17 min 24.84 sec)

But it should not take that long just to delete a few 100MB.. Ordinarily it is done pretty much instantly.. I'm not going to do anything else on the server until this is figured out..

It is also possible the server may go down if Ramnode decide just to fix the problem by moving me to a different node on the server..

[exxos]

Reply from Ramnode

I checked the node's RAID and drives and everything looks good. The RAID array is in optimal status, all drives are healthy, I/O usage on the server is very low.

and yet, when trying some wiki pages out, its maxing out the CPU and timing out now.. It runs fine on my test server and never had these problems last week

Capture.PNG (20.61 KiB) Viewed 1913 times

[exxos]

I think I've figured it out.

Ramode thought I might be running out of RAM causing the huge slowdown when dropping the database. However when I repeated the test it deleted in under a second. Though I probably rebooted the server since then. So to be on the safe side I have upgraded my server package. So we now have twice the CPU, twice the SSD space , twice the cost

However long story short, restoring the wiki database from my test server to exxoshost ultimately where things were going wrong. Another long story short...

Code: Select all

mysqldump -u root -p redacted > wiki7b.sql

Does not do what I thought it would do and just dumping the wiki database I have not worked it out yet but that function dumps the *entire* databases from the test server. Including the user accounts!

So basically when I restored that database it dropped the exxoshost user database and replaced it with the wiki one, which only had one user

It wasn't until I looked at the SQL dump itself but I saw what was going on..

Where the "mysql" table was basically where the users are stored.

Capture.PNG (41.9 KiB) Viewed 1902 times

So I used phpmyadmin to dump the wiki database and it showed this instead was what was expected ...

2.PNG (24.52 KiB) Viewed 1902 times

So after restoring that file instead, the user database was no longer trashed!

So the question is why does this not work as intended ?! Have I completely misunderstood how this function works or something ?????

Code: Select all

mysqldump -u root -p redacted > wiki7b.sql

[rmahlert]

I'd be having a few beers after all that.. Glad to see things are up again!

Off topic. 2 quotes to recover the AUN partition.

First - Minimum of $700. Estimates $1200.

Second - Minimum $250 to a max of $2800!

[exxos]

I'd be having a few beers after all that.. Glad to see things are up again!

I need a few beers after the end of each day!

Off topic. 2 quotes to recover the AUN partition.

First - Minimum of $700. Estimates $1200.

Second - Minimum $250 to a max of $2800!

Ouch Maybe do a crowd funding thing to get the cash to recover it all ? I can't even remember what all the sites were now.

[rmahlert]

I'd be having a few beers after all that.. Glad to see things are up again!

I need a few beers after the end of each day!

Off topic. 2 quotes to recover the AUN partition.

First - Minimum of $700. Estimates $1200.

Second - Minimum $250 to a max of $2800!

Ouch Maybe do a crowd funding thing to get the cash to recover it all ? I can't even remember what all the sites were now.

Same here on the beers!

Atari-users.net, TIM's MiDI site, Atariforge.org, AtariForum.com(not Atari-Forum), The Sparemint site and CVS. Highwire, DevDocs and Gokmase sites to name a few. I started AUN in 1998.

Rob

[elterwater]

So after restoring that file instead, the user database was no longer trashed!

So the question is why does this not work as intended ?! Have I completely misunderstood how this function works or something ?????

Code: Select all

mysqldump -u root -p redacted > wiki7b.sql

It's because this is the MySQL "Server" level dump command and not the individual MySQL DB Instance / Schema dump command. Because you dumped another entire server completely and then imported, you overwrote the whole MySQL "Server" with all the details (schemas, users etc) of the other MySQL "Server" (the test one). MySQL operates by having all it's system users in it's own DB schema named "mysql", so when you backed up the other test server, you took a copy of that too and used it to overwrite the important "mysql" DB.

As per the manual, you need to specify the DB name of the wiki you want to be dumping out otherwise you'll get everything including MySQL server users....

https://dev.mysql.com/doc/refman/8.0/en/mysqldump.html

Code: Select all

mysqldump [options] db_name [tbl_name ...]
mysqldump [options] --databases db_name ...
mysqldump [options] --all-databases