Importing big (and corrupt?) plaintext files into mysql broken my lamp install

Question

I was importing data from txt files into mysql (via php script). This dump is huge, it's splitted into 10 files. I imported first and second one and everything was ok - I could see all entries in phpmyadmin, etc. There were about 700 000(?) of entries. When i started to import third one something weird happened - import script loaded forever. I left it to complete its job but after veeeeery long time (and since file was the same size like previous pieces i thought it shouldn't take so long) i clicked 'stop' in browser to stop my script. Then, I tried to access phpmyadmin... loads forever and never stops. Tried to load my site (it's on my local apache2)... loads forever.

/etc/init.d/mysql restart
Stopping MySQL database server: mysqld

...also takes forever and never completes. It never actually stops it. I'm able to successfully restart apache, though.

I do have backup (snapshot - it's vm), but I'm trying to figure out what's happening. Why whole apache reacts this way? What happened? How do I prevent it from happening next time? I can revert to snapshot, import valid pieces 1 & 2 and then try 4th one... but what if 4th will break everything too?

I'm using Debian Lenny, Apache 2 and i'm accessing my site via 'virtual' domain configured in /etc/hosts and apache config file (yes, i tried accessing it via 127.0.0.1 too... doesn't work either).

Answer 1

Sorry for posting it as answer, but for some reason after i registered (i posted this question from guest account) i can't comment on my question anymore.

I tried this:

killall mysqld
debian# /etc/init.d/mysql restart
Stopping MySQL database server: mysqld failed!
/etc/init.d/mysql: ERROR: The partition with /var/lib/mysql is too full! failed!

Answer 2

There could be several things happening here. Find some more clues with the following-

• make sure you haven't filled your var partition. Issue 'df -h' and see if anything is 100% full.
• check the mysql error log- likely located in /var/log/mysql.err or something similar.
• You can always see what's going on in mysql by issuing 'mysql -uroot -e "show processlist"' You'll need to change the "-uroot" part to a valid user depending on how you have mysql setup ( "-u username -p ")

Also, killall isn't a great way to shut down mysql. Try 'sudo /etc/init.d/mysqld stop' or 'mysqladmin -uroot shutdown'. In general, try not to use killall. If you must kill, try "kill [mysql's pid]" or "ps -ef|grep mysql|awk '{print $2}'|xargs kill"