Rodrigo Barbosa wrote:
What are you running out of CentOS plus? MYSQL 5? PHP 5? Postgresql 8? That is all in Centos5. Unless you need the XFS stuff or JFS or the hardware additions, Centos 5 might have what you need.
Firewire drives containing resiserfs filesystems?
<flamebait> Why in hell would anyone still be using raiserfs these days ? </flamebait>
At least that is what I hope you mean. A 'resiserfs' is disturbing in so many different levels ...
Yes, typo, but perhaps Freudian ... It's a backuppc archive that's been running for years (since before other filesystems knew how to create files quickly) and has millions of hardlinks. The last time I tried to copy one with a non-image method I gave up after three days - and that was a smaller drive. This is a 250 gig drive about 75% full holding about 725 gigs of raw data before backuppc's linking and compression operations.
I have a 3-member raid where 2 are internal IDE drives and the 3rd is a set of external firewire drives that I periodically connect and sync (takes a couple of hours) and then rotate offsite.
If there is a better way to do that these days with LVM snapshots or something, I'd consider changing the approach even if it means starting over, but I need some way to do an image copy with no downtime - or just a momentary unmount of the filesystem so the copy is clean.