On Mon, Mar 31, 2008 at 2:32 PM, Greg Brooks gregb@west-third.com wrote:
Hi all,
I posted many weeks back about a server I've got that's incorrectly reporting available space. I'll spare everyone the rehash except to say: I'd like it fixed, and I'd like to pay one of you smart folks to do it.
Scope:
. SSH in . Figure out why a drive with 250 gb of space only shows 125gb or so . Figure out if it can be fixed non-destructively (i.e., without removing all the data on there now) . If it can be fixed non-destructively, make the fix
Happy to pay via PayPal. Interested parties hit me off-list with a rough estimate.
Thanks, Greg Brooks gregb@west-third.com _______________________________________________
Ok-I need a bit more information to see what we are dealing with . You use the term "server" which to most of us implies there are remote users. Is this a situation where downtime is a major deal killer? As in "mission critical?"
If those are both "Yes"? Then the size reporting issue is in itself less of an initial concern. The bigger risk is if you have a multiuser and/or or mission critical server with no graceful failure path in place. Perhaps there is a Linux workalike to the old Novell SFT systems? We do have s few Novell group folks on the list who could comment on that.
In a past situation where mundane degrees of reliability would do we used a hardware mirror device called Duplidisk. It's a "dumb" but hot swappable in some cases- drive mirror with a fault buzzer and some remote console in the more $$ versions. The beauty of the Dupli devices is that a single drive hardware failure becomes a less stress event. "Press Buzzer silence button" Use key and swap drive dray- relock drive tray and mirror rebuild proceeds in background. The paranoid could retire the former "good" drive to an archive copy device after the mirror has rebuilt. Or is that actually NOT paranoia bur damned good sense?
You have to do risk Vs $ evaluations as missing the call gets much more $$$ very fast these days.