Snap Server 4500 runnin Guardian O/S

[IronMike]

I hope someone can help, I am in need of serious assistance. I have a Snap Server 4500 Running Guardian OS version 2.4.040 (once we purchased the Snaps new, I have (7) 4500's, we never messed with them. I went into my Snap this morning over the network and it stated that "there are no Shares to publish" AAAHHHHHHHHHH. It was the default RAID 5 set up with 4x250 Maxtors. I looked in the event log and it stated that 2/4 devices failed and the RAID shows as failed. I can log into the Snap with no issues.

The issue is that all four drives spin and I can see all the partitions on the drives (I used forensic software to view the drives). So, I am assuming it is the controller on the Snap server that is the issue and not the hard drives themselves. How do I get the data back? Can I clone the drives and put them into another 4500?

I tried to update my Snap O/S, but I am out of the authorized time period for downloads. Any and all help is GREATLY appreciated.

[IronMike]

I apologize I forgot one thing. Can anyone provide me with the information regarding the RAID 5. I have not changed anything since reveiving it from the factory. I need to know:
Where drive is 1? Stripe size ( I assume it is 64k, not sure)? Parity distribution algorithm (Left symetric, right symetric, left asymetric, right asymetric). I have looked everywhere and can not find it. Thanks.

[blue68f100]

OK, I do not know what the strip size is. Snap has never used anything off the shelf. Contact Douglas at Snap-tech he runs a recovery service and make you a good deal. 1-866-279-2985

Maxtor's, yuck, the worst drives in the industries. 1 could have failed 1 mo after you started it up if not sooner.

The GuardianOS is running on Redhad interprise. Not sure on the early ones, some say Gentoo Linux, not sure.

V2 make these units 5 yrs old?

Look at the log and see if you can determine if you lost both drives at the same time, or 1 failed a year earier and the unit was running in degraded mode.

The Gardian units have sever features that make recovery easy if you turned them on. SNAPSHOTS , is part of the diaster recovery.

If you can make a image clone of the drive, the drive has not failed. Like you suspect, hardware failure. You should be able to view it under linux.
If you did not start up snapshots or did the steps for diaster recovery and do not have the AV active, You can move the drives (clones) to another 4500 that does not have these active. If active you must disable these features. Drive info is saved some where and will detect different drives, the reason this stuff must be disabled. The drives are numbered left to right. Take a sharpy pen and mark the drives. You do not want to get the drives mixed up.

You can monitor the boot process by connecting a monitor, or the full process using the com port (19.2k, n, 8,1 I think) and a null modem cable with terminal.

A few thing you must know. If any of the diaster recover has been setup, you will be forced into diaster recovery mode. If AV was started and not shut down it get corrupted. If not the units will boot as normal. What I discovered is the guradianOS will not touch a drive if it sees a file system. On cold boot it will transfer the OS if a clean drive, but will not format the drive for use. So formatting must be done manually. If you hot swap it will auto format and resync, if only 1 drive failure. With 2 drive failed your set has been broken, and will require some attention.

SpinRite has helped others on minor problems. Most only run Maintance mode, if it finds something it auto switches to recovery. If the data is a must have contact douglas. If you have backups, then you have a little leeway. If there is a problem it will show in the 5% area. This is a area that the tables are located in, I think.

[Phoenix32]

Run Spinrite 6 on the drives and pray. Otherwise, write off the data or pay a recovery service.

Replace the drives (with something other than Maxtors). ALL OF THEM!

Upgrade your OS.

Keep redundant backups (for business needs).

[IronMike]

I cloned the drives and placed them into an identical Snap 4500 and I get the same message in the event log. "Failed to .....raid set md0". Can it be a corrupt OS? I called adaptec and they told me my server was out of warranty. I understand that, but I should be entitled to some type of OS upgrade from Guardian OS 2.4.040, shouldn't I?

I will attempt to dial in from hyperterminal. Does anyone know the commands I should run for either diagnostics or repair? Any help would be appreciated.

[Phoenix32]

Originally Posted by IronMike

I cloned the drives and placed them into an identical Snap 4500 and I get the same message in the event log. "Failed to .....raid set md0". Can it be a corrupt OS?

Anything is possible, but it looks like a blown RAID 5 arrary to me. Two drives failed in a RAID 5 means a dead array.

Originally Posted by IronMike

I called adaptec and they told me my server was out of warranty. I understand that, but I should be entitled to some type of OS upgrade from Guardian OS 2.4.040, shouldn't I?

Nope, the software warranty is very limited. You will have to buy one of their service contracts at $600 and up for an OS upgrade. Either that or find a kind soul who already has it, but then you are talking copyright violations which we don't do here on the forum.

Originally Posted by IronMike

I will attempt to dial in from hyperterminal. Does anyone know the commands I should run for either diagnostics or repair? Any help would be appreciated.

Dude, it's not going to help. Face it, the RAID 5 arrary is blown. Without a recovery service, that data is in bit heaven. Sorry to be so blunt, but facts is facts.

David? Do you agree?

[blue68f100]

You will see the complete boot sequence and diag as it boots when monitoring throught the com ports. I do not know of any cmds. Normally when you loose 2 drives you have to use special recovery software to rebuild the tables. The reason for recovery services. If only one drive failure, hot sway would auto start the rebuild. But with 2 you do not have the needed parity info required.

I agree with Andy, recovery service is your only option if you must have the data. If not try SpinRite in recovery mode first.

Snap only provide support for the OS suppied for 3mo to 1 yr. After that they require a service contract for support. This is the main difference between Adaptec and SnapAppliance. Snap would give you free email support, where Adaptec may or may not depending upon the request.

Did you have your units sending snmp traps or email alerts? Not sure what was in v2. You really need to upgrade to v4, but do not till the raid 5 is back. Or if you are going to replace all drives go and and see if the unit will come up clean and put it back online. Then upgrade to v4.