Posted on 05-25-2010 07:18 AM
So,
What is the best practice or method if you have to down all of your casper share points to upgrade more HD space? Currently I have two drives in every server running RAID 0.
Thanks
Tom
Posted on 05-25-2010 07:41 AM
Hi Thomas,
Did you mean the Xserve is booted off RAID0? I found three Xserves at one client shop set up that way. I raised flags about it since we're required to RAID1 all boot drives on critical servers. So all 3 Xserves had hardware RAID cards with failing batteries (not sure it matters since there's no parity data to cache). An Apple tech is there now replacing the RAID card batteries.
I'm trying to sort out how to get the Distribution Point over to Windows shares, but the problem is going to be (from what I've read) syncing the Master Distribution Point with the Replica Distribution Points. I'm reaching out to the Wintel guys to see if this is something they can script (along the lines of rsync). Once we move the Distribution Points to Wintel, we can return these Xserves to RAID1 plus spare config.
PS, for those of you moving Distribution Point to Wintel, I presume AFP3 is needed and you're all using something like ExtremeZ-IP (GroupLogic)?
Don
Posted on 05-25-2010 07:42 AM
We've found the smb implementation I win 2008 server to be quick enough as a replicated distribution point...
Ben Toms
IT Support Analyst GREY Group
The Johnson Building, 77 Hatton Garden, London, EC1N 8JS
T: +44 (0) 20-3037-3819
Main: +44 (0) 20 3037 3000 | IT Helpdesk: +44 (0) 20 3037 3883
Posted on 05-25-2010 07:47 AM
Hi Ben,
Thanks for the feedback. I guess I must have had a brain f at rt...just noticed the option to use SMB for Distribution points:
JSS > Management > Distribution Points > Edit Server > File Sharing > Connection Type [SMB]
PS, Thomas...my apologies if I hijacked your thread. Was hoping to keep my comments germain to your post... :)
Don
Posted on 05-25-2010 07:47 AM
Sorry guys
No the servers are booted from a single SAS drive and all data is stored on the RAID 0 for the casper shares. I want to down all servers, plus the JSS and upgrade to my 1TB storage so I can bring them into OD and add the SUS service
Posted on 05-25-2010 07:52 AM
I suppose you could replicate to an external drive.. set that as your master distribution point.. then rebuild/image whatever it is you're doing on you JSS & then replicate to the new permanent home...
Miles? You lurking?
Ben Toms
IT Support Analyst GREY Group
The Johnson Building, 77 Hatton Garden, London, EC1N 8JS
T: +44 (0) 20-3037-3819 |
Main: +44 (0) 20 3037 3000 | IT Helpdesk: +44 (0) 20 3037 3883
Posted on 05-25-2010 07:55 AM
I have a back up on a FW drive already, I was just going to destroy the
RAID, toss in new drives, create new RAID and copy/paste the folder back
over and reapply POSIX and ACLs and call it a day, but wanted to check
if that was not the best way of doing it first
tom
Posted on 05-25-2010 07:59 AM
Not that I know off..
Ben Toms
IT Support Analyst GREY Group
The Johnson Building, 77 Hatton Garden, London, EC1N 8JS
T: +44 (0) 20-3037-3819 |
Main: +44 (0) 20 3037 3000 | IT Helpdesk: +44 (0) 20 3037 3883
Posted on 05-25-2010 08:19 AM
Tom,
I just did your method but used Carbon Copy Cloner (google it) to clone, in target-mode, my JSS machine. I then upgraded the hardware, set a new RAID array, then target-moded the JSS machine, now with new drives, and restored. Server setup just like old one but now three times the storage. Whole process took less than three hours and only 30 minutes of my actual time.
Just my two cents,
Tim Winningham
Systems Manager
The Ohio State University
Posted on 05-25-2010 08:23 AM
Thanks Tim,
I am not removing the OS drive, just the data drives so copy/paste
should do it. I think I am gonna get started on this. Thanks for
everyone's input.
-Tom
Posted on 05-25-2010 10:35 AM
well my test run is not going well, whole server is down now. Tried
upgrading a distribution point first...though I think it may have a
hardware problem