backup the entire server

[Aleksandar Milivojevic]

Rick Stevens wrote:
> As Aleksander says, block-level backup can also be accomplished by using
> RAID-1 and using the secondary drive(s) as the backup media.  You can
> then "fail" the secondary drive(s), pull them out, replace them, and ask
> your RAID to rebuild the secondaries.  This is very, very clunky and I
> do not recommend doing it--but it is an option.

Actually, what I had in mind were Solaris boxes that have some extra 
features (such as offlining, three-way mirrors, and file system locks). 
  And certanly nothing like failing the drives or ripping them out ;-)

Something along the lines:

# /etc/init.d/my-apps stop
# lockfs -fw /mount/point
# metaoffline d10 d11
# lockfs -u /mount/point
# /etc/init.d/my-apps start
# ufsdump 0f /dev/rmt/0n /dev/md/rdsk/d11
# metaonline d10 d11

If d10 was two-way mirror (the only kind of mirror available on Linux), 
redundancy is lost during backup.  If it was three-way mirror (Solaris 
has support for it), redundancy is not lost (but you need 50% more of 
disk space compared to two-way mirroring, oh well).  Plus, example above 
uses offline/online commands (not dettach/attach).  When d11 is onlined, 
only changed disk blocks are synced (not entire partition).  AFAIK, 
Linux lacks support for lockfs and online/offline operations.  So for 
large partitions on Linux this isn't a good option (huge 
detached/reattached partitions can take hours to sync on Linux, while 
offlined/onlined partitions on Solaris can sync very fast, depending on 
how many writes were to the disk while submirror was offlined).

-- 
Aleksandar Milivojevic <amilivojevic at pbl.ca>    Pollard Banknote Limited
Systems Administrator                           1499 Buffalo Place
Tel: (204) 474-2323 ext 276                     Winnipeg, MB  R3T 1L7