The other day, my mam informed me that my website wasn’t working. She was right: it was showing an error message: “Error establishing a database connection”. Oh noes.
I googled the message; one result told me check whether the error occurred just on the front end or both the back end and the front end. It was both.
I rummaged around the internet further, and discovered I could start the database up again by typing
sudo /etc/init.d/mysql start
at the command line while sshing to the server. This fixed the whole problem — or so it seemed. A few hours later, when I looked again, the error message was back. I left it for a while.
Just now, I had another look at it. I restarted the database, as above, and then tried the repair database method, while the database was actually up and running. The results were that the database was all fine. I thought it would be: I had already used Sequel Pro to check it for myself. I should have maybe tried to connect to it before I restarted it, just to see. Oh well — I’m sure there’ll be another opportunity in the future.
The advice given on a Digital Ocean community page is to check the log file. Unfortunately, it doesn’t say how to find the log file.
I found the answer on a superuser page: do some complicated-looking greppy thing:
sudo ls -l `pgrep mysqld | sed 's@.*@/proc/&/fd/@'` | grep log
Note I had to use sudo, whereas the web page didn’t mention it.
Then, once I’d found it, I had to look at it. I made up a command to do this, which surprisingly enough, worked:
sudo view /var/log/mysql/error.log
Sadly, this opened Vim, which I dislike as much as I dislike Emacs. I copied and pasted the text into a proper text editor, then tried pressing Ctrl-C to get rid of Vim. It told me to type :quit followed by the Enter key. That seemed to work.
The most useful lines in the error log are perhaps these:
InnoDB: The log sequence number in ibdata files does not match InnoDB: the log sequence number in the ib_logfiles! 160613 9:08:00 InnoDB: Database was not shut down normally! InnoDB: Starting crash recovery.
However, I don’t know what what that really means or what caused the database to shut down abnormally. Le sigh.
I think I’ll just not worry about it unless it happens again. Hopefully, it won’t.