Partition recommendations under RAID 5 in Intel Xserve

So I am preparing for a new OS X server installation on an Xserve Intel with (3) 2 TB volumes to be setup in RAID 5. Our budget would not allow the addition of an SSD drive. I would like to have two volumes under RAID 5: (1) our boot volume with OS and (1) our data and sharepoints. It would be nice to keep the OS separate from my sharepoints in case I need to do a reformat and reinstall.
Our current server setup has a RAID 1 mirrored boot volume on a G5 with an XServe RAID 5 array (all one volume unpartitioned). I also have a second FireWire hard drive connected that I use to maintain a bootable clone of our boot volume.
On our new Xserve Intel with RAID 5, is there a way to have (2) partitions under RAID 5: (1) a 150 GB boot volume and (1) an approx 3.85 TB volume that will contain all of our sharepoints. Since our server is still in transit, I do not have a machine on hand I can experiment with.
Would you create two partitions on each hard drive then join each to their respective RAID 5 array so you have two volumes—one for the OS and one for the sharepoints?
Or, with RAID 5, would it be better to keep the array as one single volume and keep the OS and sharepoints all on the same volume without any partitions? My preference would be to keep it as stable a setup as possible? Obviously, if we were to keep the OS and sharepoints as one volume, I could image the OS installation once it is stable (and without the sharepoints) and keep the image on hand just in case I needed to revert to it in case of a problem.
What would you recommend?
Thanks!

Yes, you can partition the drives as you want. The only practical way to do this is to use hardware RAID to create the RAID 5 volume with all three disks, then use Disk Utility to split this array into two partitions.
You're not going to gain much benefit from this, though. The main advantages of using different volumes for OS vs. data is for performance reasons, but you lose that here since a) you're using RAID 5 and b) all your data is on the same physical disks anyway.
About the only thing you gain is a guard against the data volume growing to the point it consumes all space on the drives and hoses your OS (because the OS is on a different volume) or vice versa, but that's a small benefit IMHO.
As for disk failure, you're no better off with partitioning than you would be with a single partition for both data and OS - either way a single disk failure will be handled by the hardware RAID card until the disk is replaced, and all your data (both OS and Data) is equally protected/at risk either way.
So my recommendation would be to use RAID 5 across all disks and be done with. No further partitioning is necessary. It is, however, your server, and you can run it however you prefer
As for the stability issue - don't bother with a disk image - it'll be outdated before you blink and will be a PITA to keep up to date.
The easier option is to create a directory for all your sharepoints and just manage your backups such that this directory is excluded from your OS backups, and included in your data backups.

Similar Messages

Intel Xserve + Xserve RAID + SCSI

I have a requirement for an Intel Xserve system that I can connect an Xserve RAID and SCSI tape autoloader to. It seems that Apple won't let me configure a system with a SCSI card and a fibre channel card in it at the same time, and repeated phone calls to them have yielded precisely nothing (very disappointed with the lack of professionalism and knowledge displayed by the Apple sales and tech staff on this matter).
Can I have such a configuration? It seems that I can only fit a fibre channel PCIe card in expansion slot 1 as it is too long to fit in slot 2 which is limited to 6" long cards. I've seen that Atto do PCIe SCSI cards but there is no mention on their site about whether they will fit/work in slot 2 of an Xserve and emails to them have gone unanswered.
If anyone has any info on this, or could point me in the direction of the info, I would be very grateful. Or if what I want isn't possible, ideas on how to back up a 7TB array without direct attached tape storage would also be appreciated.

Actually, having just looked at the store, Apple
don't seem to offer any PCIe SCSI cards, just a PCI-X
one with a PCI-X riser. Would this configuration stop
me using a PCIe card in slot 2?
Avoid the Apple BTO SCSI card like the plague. It's junk.
It's really a rebranded LSI Logic 22320. Ours never worked
right in our Xserve G5; we replaced it with an ATTO UL4D and
all of our problems went away. The right card for the intel
Xserve is the ATTO UL5D or UL5D low profile (both are PCIe).
Link is here:
http://www.attotech.com/hostadapter.html
The Apple-rebranded LSI Logic 22320 sits on a shelf gathering
dust.
Russ
Xserve G5 2.0 GHz 2 GB RAM Mac OS X (10.4.8) Apple Hardware RAID, ATTO UL4D, Exabyte VXA-2 1x10 1u

Intel xserve with hardware raid. Volume not mounting but drives show good

So this morning we came in and found that our intel xserve with the hardware raid is not booting.
Booted from OS X Install DVD and Raid Utility shows the Raidset good and all drives good.
Open disk utility and no volume form the raid will show up.
ANY IDEAS!??!

New information.
So out of the blue the Volume decided it needed to run an initializing run.
Apple care says I have to just wait until its done and see if the volume mounts.
Very strange since everything was green (good) until just then.
UPDATE:
Next day, the initializing has finished and now the Volume says (good) but the Raid Set is now "in transition". Applecare says I have to just wait till its done and see if the volume mounts.

G4 Mac OSX 10.4 server with RAID & sharepoints - Intel Xserve & RAID?

Hey all,
so we need to migrate our G4 Power Mac running Mac OSX 10.4 server with about 60 user accounts and a RAID with about 50 sharepoints to a brand new Intel Xserve with 10.4 server and a brand new RAID.
My plan for the server part of things is to put the new server online with the old one still running, make the new one a replica of the old one, then take the old one offline and promote the new one to a master.
But what to do about the sharepoints and several terabytes of data? Is there any way to do this except manually copying the data and re-creating all the sharepoints? That wouldn't be the complete end of the world as long as all the accounts/groups come across via the replication, but it'd be nice if there's a smart way to do this.
Thanks for your help!

Never tried the replica promote so I can't say that it will work. The way I migrated from 10.4 to new hardware 10.4 server was to set the new server up with exactly the same setup, DNS, IP and so on. This requires the old server be off the network for a bit. Once it was running I would have a backup of the LDAP/OD info made directly from the old server in Server Admin. Import this on the new server and all users and passwords are back in place. Now as for the share points and the data the best thing to do is to attach the old RAID to the new unit first. Hopefully you have a way to have both RAIDs connected at the same time to the new unit since that would be best.
In the interest of downtime you could immediately add the share points back in to WGM pointing to the old RAID drives. This gives you the free time to migrated the data when you want to. Make sure you know the right settings for each protocols which you use for each share point (AFP, SMB, FTP, NFS).
Now that your up and running you can move data by whatever means you see fit to the new RAID keeping in mind you want to keep permissions in-tacked. There are many clone or sync option out there and you don't need to move all the share points at once.
The main reason I chose this method is it offers a fall back and no client reconfiguring. If something is not working on the new server you can simply plug the old RAID back to the old server and your up. Second is that if you try to run both server at the same time you would need to create a whole new identity for the new server (IP, DNS and so on). Migrate data via a Network (very slow) and reconfigure all clients to point at the new server. I've done this at least 10+ times at different locations and even before the easy 10.4 backup of LDAP/OD info. Making sure you had a clean LDAP was never so easy. All swap outs took less then an hour of down time with maybe 2 hour pre-swap setup. Data movement to a new RAID was not always the case but by using this method you can move data later which is the most time consuming process of the whole swap.

Old Apple xserve Raid 1.24 and New Intel Xserve

Guys
Quick question - i currently have an Apple G5 xserve setup with an apple xserve raid {1.24} as a file sharing server in my studio
ideally i am looking to get my xserve upgraded to the new intel version but would like to know if the old xserve raid would still be compatible with the new xserve?
any info would be super appreciated!
so7

Yes - it will work just fine. The only issue is that you'll need a new Fiber channel card for the Intel XServe since it only supports PCIe cards, not PCI-X used by the previous XServes.

Intel Xserve RAID Support?

I couldn't find any mention of RAID support on the new intel xserves. Does any have any info on them. I was hoping this generation would have embedded support (no more cards).

My first batch of Intel XServes have shipped from
China. Once they get here I'll be able to tell
whether the drive bays use the daughtercard/cable
setup that facilitates the Hardware RAID Card as used
in the XServe G5. They're due in on Monday...
You lucky (and brave early adopter) guy. It would be interesting to know the controller chip on the motherboard, if System Profiler provides that information. There are a few (non-Windows) RAID vendors out there who are doing RAID 5 via driver with certain controller chipsets. Perhaps that's a path. If done right, and with journaling, it could work and be reliable. It would be a processor hit, but with all of those cores floating around, it might be doable as a low-end alternative to an XRAID.
Russ
Xserve G5 2.0 GHz 2 GB RAM Mac OS X (10.4.8) Apple Hardware RAID, ATTO UL4D, Exabyte VXA-2 1x10 1u

New Intel Xserve rebooting under network load

We added an Xserve before the end of the year (2006), and we added an Xserve RAID to it just this last week.
During the period before the RAID was installed, the Xserve was used lightly as a file server, mostly just enough for me to move some files around the office and start to get to know OS X Server.
With the RAID attached we've started to try using the system as we intend, which is to have a large common file repository where common files can be shared out to the several clients on the GigE network.
So far the results have been disappointing, the Xserve has shut down and reboot itself a few times during periods of active use. I searched this forum and found a suggestion to run memtest and I have that tool running right now on the server. The system has 2GB of RAM installed.
Is RAM a good place to start with crashes that seem to be related to network load? How else can I troubleshoot?
I've not yet called Apple, I hoped to get more info before standing on line in the phone queue.

Wes Plate-
Something wrong there. My gut says hardware. When it
shuts down is it graceful or pow and off?
I'm not sure the exact answer. We were testing the migration to the Xserve and RAID when suddenly we noticed network connection to the server was gone. I went to the server room, the display on the Xserve was off and the two raided system drives were busy for a couple minutes, then the system rebooted.
Did you enable any other services when you attached
the RAID? Are you doing anything radically
different.
I think the RAID is the only change. I've been reading up on how to do Open Director, DHCP, DNS and that stuff, but I don't know enough yet to actually do them.
Odd that nothing is in the panic log. Do you have
redundant power supplies? One of those heading south
and no backup PS could cause an immediate shutdown.
The Xserve does not have redundant power. Is that an upgrade I can do myself?
Memtest is good for rooting out errors, but you
should run it at east 10 times or so just to be
sure.
I just started a 10 times test.
Later today I'm going to test getting the RAID out of the configuration. We are trying to replace two firewire drives connected to one of our systems with shares on the RAID so everyone can access the files, so I plan to connect those firewire drives to the Xserve and then share them from there. Seems like a good test.

Intel Xserve won't boot from 3rd party drive in ADM

Right, I've got a Late 2006 Xserve Dual 2.66 with the normal non-RAID SAS/SATA board, which has been running fine for 2 years with a pair of non-Apple Seagate Barracuda ES 250GB drives (Bay 1 & 2), and an original 80GB ADM in Bay 3 as a spare boot disk.
Bay 1 & 2 are set up using SoftRAID and are partitioned into a boot volume (30GB) and a data volume (200GB), both RAID 1 mirrors. The boot volume is cloned nightly to Bay 3 (single volume, set up using Disk Utility). Data is backed up to a utility machine, so this is arrangement lets me handle any single drive failing.
This arrangement has worked fine for 2 years, and allows me to remotely boot off Bay 3 and run DiskWarrior or other disk maintenance software on the normal boot drive without having to drive across town with a firewire drive. I originally started with Mac OS X Server 10.4, but I'm now on Mac OS X Server 10.6.4.
Where things have come unravelled is, I need to upgrade the two 250GBs to something larger. I didn't realise that the Intel Xserve is pretty fussy with ADM sleds and the model of SATA drive they will work with. I didn't have any issues when I set up the Barracuda ES 250GB so wasn't expecting any problems this time.
I bought a pair of Seagate Constellation ES 500GB drives (well 3 actually, one to keep on the shelf as a spare).
I installed them in the sleds and inserted them in the Xserve. The green power LEDs came on but the drives weren't recognised in Disk Utility or System Profiler. I then added the jumper to set them to 1.5Gb/s and then they mounted. I went through my normal process of cloning the OS to them from a standard disk image I have of a known working OS install, ran Disk Utility to repair disk & repair permissions, Disk Warrior to check everything else (my standard paranoid procedures). Set the startup disk to the new install and rebooted. No joy, just the flashing question mark folder. Reboot with the option key, and the new volume doesn't appear in the list of available boot options.
I installed the drives in a firewire enclosure, and it booted without any problems, so it's not a corrupt install, it's something to do with the ADM sleds.
So to go back to basics, I tried initialising a single drive as a single volume and installed 10.6.4 server fresh from the DVD. Same result, no boot when in installed in ADM sled, boots fine in firewire case.
Erased drive, changed formatting to APM (rather then GPT), cloned OS, etc. Won't boot when in ADM sled, boots fine from firewire case. So that screws the theory that drives have to be GPT to boot on Intel hardware.
So I ordered a set of WD RE3 500GB drives (x 3) and repeated this whole process. Same deal, boots fine in firewire, won't boot when installed in ADM. GPT vs APM doesn't make any difference.
Try #3. Ordered desktop variants rather than server class drives. So armed with a pair of WD Caviar Black 500GB. Same deal, boot fine from firewire, will not boot when installed in ADM. Also required jumpers for 1.5GB/s and also the undocumented pins 7&8 (apparently this might be related to the new Advanced Format they are using)
Try #4
A pair of Seagate Barracuda 7200.12 500GB. This is where it gets tricky. One of the Barracuda will boot the Xserve when installed in an ADM, and one will not. So 2 Barracuda 7200.12, both same model number (ST3500418AS), both same firmware (CC38), yet they behave differently.
Try #5
Barracuda ES.2 250GB (x 3). Won't boot in ADM, will boot firewire.
Try #6
WD Caviar SE16 250GB (WD2500KS)
Boots in ADM (Yay!!) and firewire
Try #7
WD Caviar SE16 500GB x 2 (WD5000AAKS)
Boots in ADM and firewire
I've managed to get access to G5 Xserve (2.3DP) so I've repeated all this testing on it as well. The G5 doesn't care about the model of drive and boots from any of them using a 10.5.6 install with APM formatted drives.
I've tried all these drives in 4 different models of ADM sled. One thing I have determined is that the model of sled determines if you need to use the 1.5GB/s jumper. I have 2 sleds from G5's, they require the jumper, a 80GB Intel ADM sled with separate data and power plugs requires the jumper, and an Intel 160GB ADM sled with the ribbon cable and combined data/power plug doesn't require the 1.5GB/s jumper. APM vs GPT formatting again makes no difference.
So on the Intel machine, the drives work in the sleds, appear as normal drives on desktop or in Disk Utility when logged in, and also appear as normal drives when connected via SSH if Auto Login is disabled. So to me they appear to be working drives.
It appears to me that something changed at a low level between the Barracuda ES and ES.2 generations of drives (for WD drives as well since the Caviar SE16 that boots is a similar age to the Barracuda ES)
So then why won't it boot from them? Is there some magic dance I've missed?
And how come one of the 7200.12 will boot, while another identical model won't?
Any tips or suggestions greatly appreciated.

I don't have an answer for you, but you deserve thanks for all the useful hardware information you posted from your trial-and-error testing.
My impression has that the XServes need specific kinds of firmware and hardware monitors on their drives in order for everything to function smoothly. My own efforts swapping drives in ADMs has produced "mixed results" to put it mildly. It would be so much easier if we knew what kinds of drives worked properly in what kinds of ADMs and on which Macs/XServes.

Intel Xserve not using full CPU

When submitting a batch to qmaster, I notice my other g5 machines utilize almost all the processor, and on the intel xserve, all 4 CPUs usually max out at 40% tops. Is this a limitation, to be expected, etc?

Things do look clean - not familiar with all of the Processes, but some of the common "culprits" do not appear. Nice job in keeping these to a minimum. Most people have 60-80 going. Maybe someone else can look down your list and find something that can be eliminated easily for NLE work. I just don't see any potential problems, but remember I do not recognize some of those.
I like the Virtual Memory setting, with only one question: you mention a partition on the RAID 0. Why a partition? The reason that I ask is that Windows first sees your 2 HDD's as one (the RAID), then it sees your partitions as several. If accessing what it now thinks are 2, or more, physical HDD's, it's telling the controller to get the heads in a couple of places at the same time. You probably have a good reason for doing it this way, but it does throw up a yellow flag to me. Having the RAID managed by the MoBo chip is not the fastest/best, but is usually more than adequate. It does save on a US$400 RAID card with multiple channels/chips.
You have provided me, and others, with food for thought. I'll keep looking over your data, but just see nothing, except for the Page File on a partition, that might affect performance.
One test that I'd run would be a slight re-config on your Scratch Disks. I'd do a Save_As, or a Save_As_a_Copy for your Project. Leave your original totally alone. Do a test with the Scratch Disks set to "Same as Project." This will put them on the RAID 0. Then, perform the same operations and monitor the time. Unfortunately, you're investing hours into this, but you might get enough of a performance increase to justify it for future Projects. One way around the time, would be to do a "test" Project with shorter versions of similar Assets. Do all processing on that Project, as configured, and time the results. Do the Save_As, make the changes to the Scratch Disks, and process again. Any difference?
Good luck, and thanks for gathering the data,
Hunt

10.4 on an Intel XServe w/ Xeon 5400

I suppose I already know the answer to my question that I'm about to ask, but I thought I'd toss it out to the forum as my final hope and prayer.
I've got a Intel XServe (3 GHZ Quad Core Xeon 5400) that came with 10.5 Server installed and as the included media in the box. I've got no problem rolling out 10.5 for the particular project I'm working on except that the project I'm working on isn't going to give me enough time to properly roll out 10.5 both to the server and clients. So, the plan was to put this new XServe in place with 10.4. I'd be able to use the same DNS name and I wouldn't have to rebind any clients for OD. The main purpose of this project was to get this new XServe in place, and so I don't mind staying at Tiger Server.
The XServe comes set up with a RAID configuration. I've deleted the RAID sets a few times using the 10.5 DVD. Unfortunately, after doing that I cannot figure out how to use the three drives as standard drives. From what I can tell, I've got no choice to use RAID.
I also cannot use the 10.4 DVD to install the operating system. Whenever I boot to that disc, none of the drives show up in Disk Utility or inside the OS installer. This happens whether I still have a RAID array set up or not.
I suppose my main question is -- is it even possible to put 10.4 Server on this hardware?
Any help or advice would be greatly appreciated. I'll monitor this thread throughout the day and respond as fast as I can with any questions.

I pretty much agree with the idea of going to 10.5 except that because mixed environments haven't worked so well in the past. And because of that, I don't much have a leg to stand on when it comes to presenting that idea to those in charge.
I think I'm just going to have to put the whole project off. I've tried just about everything under the sun. 10.4 is not getting installed on this XServe.
Thank you though.

Can I add partitions on a RAID 0 setup?

Hi everyone,
I just recently bought a used macbook pro, (15-Inch, Early 2008) that the previous owner put a RAID 0 setup on. I want to try out the OSX Yosemite Beta, but i want to try it on a separate partition so I can keep my Mavericks environment intact. I don't know i it is possible, but I would like some help on how to make a partition with the RAID setup on it. (I do have an external hard drive too, is there a way to run Yosemite off of the external instead of messing with the internal on my MBP?) Thanks in advance for your help!!!
Tommy

Camelot wrote:
This doesn't sound like an XServe RAID question - the XServe RAID is a (now-discontinued) 14-drive RAID enclosure designed for server applications.
I apologize for not knowing where. I know how forum members get upset when people use the wrong forum. It was unintentional.
For a start there's a question of the enclosure... Is the enclosure a 'dumb' enclosure that just connects the drives to your host system, or does it have a built-in RAID controller.
It is not a dumb enclosure - it does have jumpers for setting it up under different strategies.
That is going to be critical in answering your question. If the enclosure has a built-in RAID controller then you will want to use that to manage the RAID 1 mirror setup and you'll likely need to use software provided by the enclosure vendor to set that up.
Where I am at is that DU does NOT allow this system to be used and recognized by TC. I had to set the enclosure under RAID 1 and that is the only way the TC would allow it to be recognized by my network
Once that's done your host system will be presented with a 1TB volume (the RAID/mirror is hidden from the OS) and you can partition that like any other drive.
If the enclosure is dumb it's a little more complex since you'll need to configure the mirror via Disk Utility and then partition the resulting volume (note also that throughput will be lower in this mode since the OS has to write twice as much data over the link to the enclosure - with a hardware RAID in the enclosure the OS only has to write one copy of the data and the enclosure takes care of writing it to both disks).
So, what's the enclosure you're using?
I am using the Hornettek unit.
Thanks for your help. It took me a long time to get to where I am with this unit. TC evidently will not allow it to be recognized if I used DU to configure it.

New intel Xserve with Leopard Server setup assistance....

Hello
Im quite a newbie when it comes to administrating OS X Server, Xserves and RAID's so please bare with me if the following question seems obvious or dumb...
Currently I have inherited the following setup - G5 Xserve running 10.4.10 Server which has the following services running = Kerio MailServer, AFP, DHCP, DNS, Firewall, NAT, OD, VPN and Windows. This Xserve is also connected to an Apple RAID via fibre channel for file sharing and also to a Tape multi-changer backup unit via scsi.
I have just bought a new Intel Xserve and have installed 10.5 Server on it. I have this server as I would like to get iCal Server and Directory up and running and eventually share some of the G5 services. For iCal Server and Directory I assume ill need the 10.5 Server set up as Open Directory Master but my worry is I have that already on my G5 Xserve. If I was to migrate the 25 usernames from there to the 10.5 Server, would every user still be able to connect to the RAID (as the RAID is still physically connected to the G5 Xserve) for file sharing?
I hope some one could advise or recommend a better way, let me know if you need me to try and clarify the question better!
Thanks in advance

Hello
Im quite a newbie when it comes to administrating OS X Server, Xserves and RAID's so please bare with me if the following question seems obvious or dumb...
Currently I have inherited the following setup - G5 Xserve running 10.4.10 Server which has the following services running = Kerio MailServer, AFP, DHCP, DNS, Firewall, NAT, OD, VPN and Windows. This Xserve is also connected to an Apple RAID via fibre channel for file sharing and also to a Tape multi-changer backup unit via scsi.
I have just bought a new Intel Xserve and have installed 10.5 Server on it. I have this server as I would like to get iCal Server and Directory up and running and eventually share some of the G5 services. For iCal Server and Directory I assume ill need the 10.5 Server set up as Open Directory Master but my worry is I have that already on my G5 Xserve. If I was to migrate the 25 usernames from there to the 10.5 Server, would every user still be able to connect to the RAID (as the RAID is still physically connected to the G5 Xserve) for file sharing?
I hope some one could advise or recommend a better way, let me know if you need me to try and clarify the question better!
Thanks in advance

VM Server installer does not recognize RAID configuration in Intel ESRT2

Hello,
I'm having a big issue here trying to deploy my third Oracle VM Server.
I have:
- A Intel Server Board S3420GPLC
- A Intel Xeon X3440 processor
- 16GB DDR3 ECC Memory
- Six 1TB Hard Drives
What I'm trying to do:
- Create a RAID array using Intel ESRT2 (Embedded Server RAID Technology II) using all the discs (done)
- Install Oracle VM Server into the array
- Install some VMs....
The problems:
- When I boot the VM Server installer from USB stick I got an error saying that there is not enought memory. This issue is solved by typing "linux mem=15GB"
- I got a screen per disk warning me about that the discs need to be initialized and all the data will be lost
- When I go to the disk partitioning screen all the disks are listed separated, the installer simply don't recognize the RAID configured before.
What I already tryied:
- Change the RAID controller from ESRT2 to Intel Matrix
-> Didn't solve.
- Create a Driver Disk with the disk images supplied by Intel here
-> I launch the installer using "linux mem=15GB noprobe=ata1 noprobe=ata2 ....". When Im asked by the installer to load the drivers it doesn't recognize the disk, so i couldn't load the drivers.
Now, the questions:
- Is there another driver I can use do make the vm server installer recognize the RAId configured?
- Is there a better way to setup this server, considering that it needs to be "self contanied" (store and run vms)?
Edited by: xyp on 14/10/2010 11:37

So...
- I have installed VM Server into ordinary linux partitions on the first disk.
- I have created a big partition in the first disc with the space not used for the VM Server
- I have created five equal partitions on the other discs
- I have configured software raid 10 using mdadm
Now the problem seems to be solved, i have a storage repository with almost 3TB and can deploy my VMs.
Unfortunatelly only my windows 2008 r2 hardware virtualized vm is working, the linux paravitualized guests not. but this new issue will be a subject for another topic.

Intel Xserve will NOT boot from hard disk drive

I have two Intel Xserves 2.0 QC machines
I have tried to re-install version 10.4.11 OS (don't ask) onto a Hard disk drive that is in a Disk drive caddy and inserted in the machine.
When booting from The installation DVD media
The disk is seen by the "Disk Utility" (the disk is partitioned as GUID boot) and the installation proceeds to the disk.
Once the first pass of the installation is complete and the system reboots, the system shows the folder and ? mark.
If I boot and hold "ALT" key to view all avaialble boot options the Hard Disk Drive does NOT appear as valid boot disk.
It does show, the still inserted Installation DVD.
I CAN boot from an external firewire drive and the Hard disk drive is visible to the OS once logged in.
I have tried different caddy postions (left, middle and right)
I have tried bona fide Apple supplied Disk drives
I have tried various different caddies
I have tried installing Operating System 10.5.* and 10.6.8 server
I have run SMC reset and system reset
But still I cannot boot off from any Hard Disk Drive inserted in a caddy at the front of the system.
Has anyone seen this behaviour?
Is it fixable?
Do I have replace something like the Hard Disk Drive back plane circuit board? Or even the main logic board?
I have now run out of ideas and cannot find any help in discussions or just googling.
ANY help would be very appreciated
Regards
Tony

Antonio Hi !
QC = Quad core or QX as it is written on the "pull out" serial number tab at rear of system
I have 3 late 2006 Intel xserves
On my intial posting I declared 2 were 2.0 machines - when I was trying to be concise
2 of the macines are 2.66 systems and 1 is 2.0 system - all are QX machines
the 2.0 and 1 of the 2.66 will NOT boot from the ADM and disk inside the system - whereas the other 2.66 system will.
I have many different Caddies and these have different product/model number labels
603-9965
603-9545
603-9796
these are marked "intel SATA"
The 80Gb was only one of the drives tested - I have used Seagate 400Gb 750Gb and 1Tb - all of which have been used in systems at one time or another
and as I said the drives ARE visible once the operating system (either installation DVD or external firewire are booted from) - they are just NOT visible when the system is searching for boot media.
Thanks for the guide - I will look through that and continue to try and find an answer

Intel Xserves Crash only on weekends

We have 10 Intel XServes that we received in early December. For some reason anywhere between 2-6 of the 10 crash and become unresponsive, only on the weekends, usually Saturdays. Very odd.
I've been checking them every weekend since we've purchased them and it has been consistent for the last 5+ weeks. They crash only on weekends.
I found this in the crashreporter log on one of the servers, they all have similar logs:
Sat Dec 23 14:28:50 2006 crashdump[120]: crashdump invoked as panicdump
Sun Dec 24 02:11:08 2006 crashdump[1814]: crashdump started
2006-12-24 02:11:09.210 crashdump[1814] threadreadstack: stack appears to be in inconsistent state to trace. Truncating stack.
2006-12-24 02:11:09.210 crashdump[1814] threadreadstack: frame pointer is zero, continuing.
Sun Dec 24 02:11:11 2006 crashreporterd[118]: mach_msg() reply failed: (ipc/send) invalid destination port
Sun Dec 24 02:11:11 2006 crashdump[1814]: Started writing crash report to: /Library/Logs/CrashReporter/drone.crash.log
Sun Dec 24 02:11:12 2006 crashdump[1814]: Finished writing crash report to: /Library/Logs/CrashReporter/drone.crash.log
Sat Jan 6 08:26:57 2007 crashdump[113]: crashdump invoked as panicdump
Sat Jan 13 10:01:33 2007 crashdump[120]: crashdump invoked as panicdump
Does anyone have any ideas? What other logs should I be checking?
Quad G5s, Intel XServes Mac OS X (10.4.8) Xsans, XServe RAIDs, oh my...

The "drone" is our encoding software. That may be the cause, but i'm not entirely sure of course.
Here is the panic log:
Sat Dec 23 14:28:50 2006
panic(cpu 3 caller 0x00141505): zalloc: "kalloc.256" (1915824 elements) retry fail 3
Backtrace, Format - Frame : Return Address (4 potential args on stack)
0x4bc93cf8 : 0x128d08 (0x3c9a74 0x4bc93d1c 0x131de5 0x0)
0x4bc93d38 : 0x141505 (0x3ca780 0x3c9d1c 0x1d3bb0 0x3)
0x4bc93d98 : 0x12d97d (0x1945b70 0x1 0x4bc93dc8 0x12db75)
0x4bc93dc8 : 0x12e192 (0xa8 0x1 0x4bc93e18 0x140259)
0x4bc93df8 : 0x63421142 (0xa8 0xa9b2680 0x4bc93e58 0x1a16ce)
0x4bc93e28 : 0x6341c64d (0x63476c68 0x0 0x558c161c 0x20)
0x4bc93e98 : 0x6343a294 (0xa8 0x0 0x2 0x0)
0x4bc93ee8 : 0x6343ad67 (0xaaef800 0xac94000 0x18 0x634211dd)
0x4bc93f78 : 0x6341fa8e (0xaaef800 0xaaef814 0xaceb8c4 0x0)
0x4bc93fc8 : 0x19a7ac (0xaceb8b0 0x0 0x19d0b5 0xa3d6c64) Backtrace terminated-invalid frame pointer 0x0
Kernel loadable modules in backtrace (with dependencies):
com.apple.filesystems.acfs(2.7.1)@0x633c0000
dependency: com.apple.iokit.IOStorageFamily(1.5.1)@0x516000
Kernel version:
Darwin Kernel Version 8.8.4: Sun Oct 29 15:26:54 PST 2006; root:xnu-792.16.4.obj~1/RELEASE_I386
Sat Jan 6 08:26:57 2007
panic(cpu 2 caller 0x00141505): zalloc: "kalloc.256" (1829872 elements) retry fail 3
Backtrace, Format - Frame : Return Address (4 potential args on stack)
0x4bd0bcf8 : 0x128d08 (0x3c9a74 0x4bd0bd1c 0x131de5 0x0)
0x4bd0bd38 : 0x141505 (0x3ca780 0x3c9d1c 0x1bebf0 0x3)
0x4bd0bd98 : 0x12d97d (0x1945b70 0x1 0x4bd0bdc8 0x12db75)
0x4bd0bdc8 : 0x12e192 (0xa8 0x1 0x4bd0be18 0x140259)
0x4bd0bdf8 : 0x63423142 (0xa8 0xad8f280 0x4bd0be58 0x1a16ce)
0x4bd0be28 : 0x6341e64d (0x63478c68 0x0 0x558c161c 0x20)
0x4bd0be98 : 0x6343c294 (0xa8 0x0 0x2 0x0)
0x4bd0bee8 : 0x6343cd67 (0xaabec00 0xacab900 0x18 0x634231dd)
0x4bd0bf78 : 0x63421a8e (0xaabec00 0xaabec14 0xac450c4 0x0)
0x4bd0bfc8 : 0x19a7ac (0xac450b0 0x0 0x19d0b5 0xa3d6c64) Backtrace terminated-invalid frame pointer 0x0
Kernel loadable modules in backtrace (with dependencies):
com.apple.filesystems.acfs(2.7.1)@0x633c2000
dependency: com.apple.iokit.IOStorageFamily(1.5.1)@0x516000
Kernel version:
Darwin Kernel Version 8.8.4: Sun Oct 29 15:26:54 PST 2006; root:xnu-792.16.4.obj~1/RELEASE_I386
Sat Jan 13 10:01:33 2007
panic(cpu 1 caller 0x00141505): zalloc: "kalloc.256" (1823792 elements) retry fail 3
Backtrace, Format - Frame : Return Address (4 potential args on stack)
0x4bd43cf8 : 0x128d08 (0x3c9a74 0x4bd43d1c 0x131de5 0x0)
0x4bd43d38 : 0x141505 (0x3ca780 0x3c9d1c 0x1bd430 0x3)
0x4bd43d98 : 0x12d97d (0x1945b70 0x1 0x4bd43dc8 0x12db75)
0x4bd43dc8 : 0x12e192 (0xa8 0x1 0x4bd43e18 0x140259)
0x4bd43df8 : 0x799d6142 (0xa8 0xa423280 0x4bd43e58 0x1a16ce)
0x4bd43e28 : 0x799d164d (0x79a2bc68 0x0 0x558c161c 0x20)
0x4bd43e98 : 0x799ef294 (0xa8 0x0 0x2 0x0)
0x4bd43ee8 : 0x799efd67 (0xafc3c00 0xae4e200 0x18 0x799d61dd)
0x4bd43f78 : 0x799d4a8e (0xafc3c00 0xafc3c14 0xad2d8c4 0x0)
0x4bd43fc8 : 0x19a7ac (0xad2d8b0 0x0 0x19d0b5 0xa3d1e68) Backtrace terminated-invalid frame pointer 0x0
Kernel loadable modules in backtrace (with dependencies):
com.apple.filesystems.acfs(2.7.1)@0x79975000
dependency: com.apple.iokit.IOStorageFamily(1.5.1)@0x516000
Kernel version:
Darwin Kernel Version 8.8.4: Sun Oct 29 15:26:54 PST 2006; root:xnu-792.16.4.obj~1/RELEASE_I386
And here is the asl.log. I only included a few of these entries, as there are a lot there a quite a number of the same identical entries.
[Time 2007.01.16 16:31:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 16:31:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 17:01:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 17:01:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 17:31:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 17:31:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 18:01:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 18:01:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 18:31:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 18:31:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 19:01:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 19:01:36 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 19:31:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 19:31:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 20:01:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 20:01:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 20:31:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 20:31:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 21:01:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 21:01:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 21:31:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 21:31:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 22:01:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 22:01:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 22:11:14 UTC] [Facility daemon] [Sender ntpd] [PID 7931] [Message time reset +0.137353 s] [Level 5] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 22:31:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 22:31:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 23:01:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no name available via DNS for 192.168.60.69] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
[Time 2007.01.16 23:01:37 UTC] [Facility user] [Sender servermgrd] [PID -1] [Message servermgr_dns: no hostname set and unable to detect via DNS, services may not function properly - use changeip to repair] [Level 3] [UID -2] [GID -2] [Host Encoder-9]
Quad G5s, Intel XServes Mac OS X (10.4.8) Xsans, XServe RAIDs, oh my...

Partition recommendations under RAID 5 in Intel Xserve

Similar Messages

Maybe you are looking for