I had a bit of a disaster, in that I lost a weeks worth of backup, so I now have a content store which is 10 days old, and a db which is fresh, but with the old backups accidentally deleted.
Is there any way to get this working? What would be the best procedure?
Running on 3.4.d, Debian, MySQL, Tomcat
This was my status. Thanks to information from Loftux, and help from a young computer wizard, our share is now up and running again. Albeit I did loose the data of 10 days, so we have tightened up on back-up procedures.This is how we cleaned up and managed to get a 10 old data directory to work with a 24h old copy of the DB. The problem is that the DB contains pointers to the data (files, blog posts, pictures etc) in the alf_data directory, in the alf_content_url table. When it finds nothing it throws an error and renders the site unusable. You could delete the all the pointers that had been created after the date and time for which there is no matching data, but there are other tables in the database that use the information in the alf-content-url table, so this would not be any good. The first step was to get a listing of all the items in the content store## Visit the contentstore and list its contents in a text file## Also format it to be matcheable against the next file we will createcd /opt/alfresco-##/alf_data/contentstore
find . | grep - > contentstore.txt
sed -i 's/\.\//store:\/\//g' contentstore.txt
Then get the alf_content_url table listed in a similar way## Dump the alf_content_url table to a text file (with nice linebreaks)mysqldump -u YOUR_DB_USER -p YOUR_ALFRESCO_DB alf_content_url | sed 's$),($),\n($g' > alf.sql
Then compare the two, to see what is in the DB that is NOT in the alf_data contentstore.## Script to get a list of the missing filesdblist=( `cat "alf.sql" `)
filelist=( `cat "contentstore.txt" `)
left=${#dblist[@]}
for pointer in "${dblist[@]}"
do
exists=false
pointerpath=( `echo $pointer | cut -d "'" -f 2 `)
for filepath in "${filelist[@]}"
do
if [ "$filepath" == "$pointerpath" ]; then
exists=true;
fi
done
let left-=1
echo $left
### List files that does not exist
if [ "$exists" == 'false' ]; then
echo $pointer >> pointers.txt
echo $pointer
fi
done
Then we need to create something for the pointers to point at, a dummy file. The start of file and end of file "{" and "}" ## Missing files are now listed in pointers.list create a dummy file containing only {} and place it in the contentstore (create a path similar to the ones already there)## Putting {} in the dummy file avoids problems with missing files containing JSON-data.## Lets say we find that all content files over id 7000 are missing, point them to the new file we created with the SQL query belowUPDATE `YOUR_DB_NAME`.`alf_content_url` SET `content_url` = 'store://2011/11/30/9/26/713d9b34-3f6b-472c-8b5e-4e4b74a5b66.bin' WHERE `alf_content_url`.`id` >7000;
We then copied in the alf_data directory (10 day old back-up), set the reindexing index.recovery.mode=FULL
started alfresco and crossed our fingers. Success. I then did a search from inside share to delete any files that were just "{" and "}". And we now have a working site. I have seen, somewhere in the forums, a method to get a working site from a DB back-up that is older than the alf_data directory. Here is now a method for the reverse, a working site from an alf-data directory that is older than the DB back-up.Do your back-up and thanks to all that helped with this!
