Broken SW-Raid / HDD and Rebuild fail

  • @Smallserver Server are from hetzner too, have there some spare/backup server so thanks.
    @biologist, thanks for your tipps. It was a fast server install and restore from a old one. so that was quick and dirty. Yes this is the third broken 3TB HDD by them ...

    I think the reinstall is the best solution, but im a bit afraid of the long work ... With a few customer and so on .. no problem ... used imapsync too in the past
    but here at over 120 customer ... ouman....

  • Just start, if it takes time, it will take time :-)
    In all the years at Hetzner I never had a broken HDD. Maybe I shouldn't have said that .. nock nock nock.

  • Just start, if it takes time, it will take time :-)
    In all the years at Hetzner I never had a broken HDD. Maybe I shouldn't have said that .. nock nock nock.

    The problem is just this one hdd-model... it's a mess!
    The first hdd broke after one year (sdb). Fortunately, this one was replaced immediately as just one week later sda said goodbye. While sdb was replaced by a new hdd, sda got exchanged by a used one that had - at least as of smartctl-log - "a history". In other words: obvisously, this one was not 100% ok! Just three weeks later, it had to be replaced again - this time by a NEW Toshiba-drive. Works so far...

  • had this exactly the same ...

    Is there any recommendation on whats the best way for backup and restore ? I did this serval times ago but with only a bunch of customers ;-)

  • @FraggDieb

    Hello ;

    • You must just rsync some directories (/var/www/virtual, /var/mail/virtual, /var/www/imscp/gui/data, /etc/imscp) on a remote (backup) server
    • Dump all your db and upload them on the remote server
    • Install new system (with same i-MSCP version and same services)
    • Restore all databases
    • Rsync the directories from remote host to new system
    • Rerun installer after changing base server IP if needed, this will recreate all unix users and fix permissions
    • Maybe some other adjustments...

    I've processed this way for a big server some time ago ( a server from @c0urier )


  • It's always a good idea to make those backups everyday. I use automysqlbackup for my databases. I don't have rsync running, but I am planning something with lftp for all else (my backup storage is a NAS, doesn't have ssh). I make backups manually every day to secondary storage. Last week however, something went wrong with that and my hdd filled up :-)

  • Hi,

    thanks @Nuxwin how long does it takes for you to do such work? How long does it took you for the @c0urier server and how big was it? I think thats a good Idea but with downtime...

    My first idea was like: copy partitions to new HDD, format them and copy all files while online. after that shutdown and boot in recovery modus, mount all partitions and copy last files like /proc /tmp /dev /sys so now everything has to be on the new disk but here starts my problem, what I have to do, to boot from the new disk ... ? I think when this is done ... server boots up from new disk, everything is good :) ... Im at the point to reboot but ...missing link ;-)

    @Nuxwin ... Other question. Would it be possible for you to do this transfer for me and how can i pay you to do this? I could do the first steps, rsync the directories and database to other Server, I think this will take 2-3 hours, the reinstall is no problem but the import of the backups ...

    @smallserver: We copie all the customer backups one time per week to another server. its a small script i wrote to backup everything in /var/www/virtual/*/backups