We recently performed a migration and upgrade on our moodle site but somehow seem to have lost the mysql db. We have all the hash files from moodledata but are unable to use those files. Is there any way to perform a reverse db build that will organize the hash files appropriately?
Thanks in advance for any help.
Dan L
By 'hash files' I assume you mean the the 'moodledata' files area.
If, as it sounds like, you have lost the complete database then the files stuff represents only one of hundreds of tables that go to make up the Moodle database. If you have genuinely lost your database with no backup then you have lost your site.
Sorry to be bearer of bad news.
Thanks for response Howard.
You are correct in thinking I am referring to the files in the moodledata folder (completely in tact).
What happened is that an incorrect moodle folder was backed up (an archieved version rather than the most recent version).
Is there anything that can be done with the files in moodledata folder to aid rebuilding a new site?
Thanks again.
Absolutely nothing.
Those files are only the stuff that was uploaded. Absolutely everything else was in the database. Even if you take the view that you'd like to recover the files you'll never do it. Again, all the information you need was in the database.
It's bad news I'm afraid.
Useful depends upon if teachers backed up their courses, if you had automated backups going to 'filearea', etc.
If you can find files that are gzip in the sea of files in moodledata, they might be backups of courses.
From moodledata/filedir/
find ./*/*/ > hashes.txt
hashes.txt file will be large and lines will look like:
./00/00/
./00/00/00007e0b20e39c78158bbc1268309eec06888b0a
./00/00/00000fc42d94251665e3f2ca6511fdaa352080f3
./00/03/
./00/03/0003bcd231a6ee0d78b8df313643930b566b8d12
Create a bash shell script called possibles in moodledata/filedir/
#!/bin/bash
#
rm possibles.txt
touch possibles.txt
for i in `cat ./hashes.txt`
do
echo "Hash in que:" $i >> possibles.txt;
file -b $i |grep gzip >> possibles.txt;
done
echo "Done!";
That will create a possibles.txt file ... a large one - and you'll have to look through it to find what could be backups.
In possibles.txt you, hopefully, will see entries like this:
Hash in que: ./07/4f/074f9c58068de1f24902744625386c577c40663a
gzip compressed data, last modified: Tue Apr 4 06:40:11 2017
And even if you can't see a name, one could copy them out of
moodledata/filedir/X/X/X to a holding directory (/possibles/) by names
you created ... pbackup1.mbz.
cp ./07/4f/074f9c58068de1f24902744625386c577c40663a /possibles/pbackup1.mbz
Then using tar to view the contents of the .mbz files you've named (may not be backups) looking for moodle_backup.xml.
tar -tvf pbackup1.mbz |grep moodle_backup.xml
-rw-r--r-- 0/0 5558 2018-09-14 08:56 moodle_backup.xml
If last command shows there is a moodle_backup.xml then one could extract just that file to find the original name of the course.
Pain in the arse. Very time consuming. But, you have something ... which is better than nothing - IMHO.
'spirit of sharing', Ken
Thanks to everyone for their interest and suggestions. I'll give Ken's idea a go and see what I can find.
Cheers.
DL
I've managed to recover a good number of courses using some of the info Ken shared. Here's what I had to do for those who might find themselves in the same boat one day.
First ... I did not have access to the server so I was stuck working only with a backup of files in moodledata\filedir so I created a standalone server of my laptop to work from.
Second ... I used treesize to determine large files within moodledata\filedir -- very tedious but managable.
As Ken mentioned, I looked for the larger files which might actually be backup files.
Once I determined the larger files I simply dragged them into a bogus course and restored them ... if the file was actually a course backup file moodle would simply restore the course (no need to change the course extension to zip or mzb this way) ... if the file was not a backup moodle would reject the file.
From there I simply backed up the file (course.mbz) and then restore to the live site.
Thanks again for everyone's help on this.
Cheers.
Congrats! There is usually more than one way to 'skin a cat'.
Now folks have a Windows way as well - even though it's a little 'harder' ... IMHO.
Needless to say ... but will remind ... bet you are working on backup of database on a daily basis, huh? ;)
'spirit of sharing', Ken
'Necessity the mother of invention'???? But thanks!
'spirit of sharing', Ken
