Snap Server 2200 Error Log

[Mikey_Snap]

I'm attaching part of my error log from my Snap Server 2200 - the errors started on April 23rd but there's a forum limit on the size of the DOC file you can upload so can only get the first two pages (out of 10).

Specs:
Model Software Hardware Server # BIOS
2200 4.0.860 (US) 2.0.4 824193 4.0.855

How do I interpret this? Is the motherboard going or one of the mirrored drives?

[blue68f100]

Sorry for the delay, I have been out of pocket with limited access.

I get an error on the file, but was able to read it. With the logs save as HTML.

The error accours every time you boot the server by the log. This error is normally an indication that you may have a HD failing. My suggestion is to make sure you have a current backup of your files. I would then mark position and jumper settings, remove the HD's and see if you can make a copy of the SnapOS Boot trackes with Digital Dolly. In most cases you do not need it if your running raid1, but recommended. If Maxtors HD's I have seen these fail with in days of each other.

Do not let Windows see the HD's or you will loose everything. Both Spinrite (CD) or Linux OS will not modify the HD's. Neither can read it, so only use the Boot CD for Spinrite or Linux OS to run DD from terminal mode.

Then I would run SpinRite (in maintance mode) on the HD's to see if you are experiencing media problems. Spinrite works at the HW level and does not attempt to inturpit the file structure. Do this on both HD's. If SpinRite has a problems in the 4-5% area, the HD is failing. If you have Maxtors HD's, your on borrowed time, go ahead and replace the HD's with something good. Use the guide in the FAQ's. We normally recommed the WD Raid Edition (RE) drives.

I do not recommend start and stopping the Snaps on a daily bases. This puts a lot of stress on the power supply. And this load increases as the HD get old and start failing. Use a UPS and run it 24/7. If you want auto shutdown you need to a APC Smart-UPS with a network card.

[Mikey_Snap]

I added another HTM log file (zipped) from the debug screen (ran a "info log p-1").

Wasn't planning to power it down daily, was forced to because it would lockup, go into panic mode, and then no one could access it over the network.

I replaced the UPS on Friday afternoon. Then ran two check disk routines in a row. Is it worth the troubleshooting steps you described, or should I just backup the data and swap drives?

I think I'm 6 months past the 3 year warranty. So two new drives would run roughly $200 - or a new Snap Server 210 (assuming it's the power supply or motherboard) is $1200. Any idea what Adaptec charges for repairs?

[blue68f100]

Adaptec has dropped support unless under warranty. So I dought they will do any service to it. They will push for you to get a new server. The newer 210 runs the more robust GuardianOS. This is a Redhat Interpise release w/ the Guardian frontend.

You can try reloading the SnapOS, and see if it's due to corrupted OS, not likely.

I really think your HD's are going south. Maxtors have the highest failure rate in the industry, and I would not trust them with anything. I have seen their drives fail literally days apart not giving you time to get replacements. This may be happening NOW. So backup while your unit is still working, it may not tomorrow. You may be on borrowed time. I had a refurbished 2200 that have had bad drives, adaptec did a poor job in testing (none) the unit. Snap Appiance did not use the cheap Maxtor drives, it's a Adaptec thing to save money and get more support work. Using Spinrite v6 would confim this. And may even beable to repair the drive. This is an execelent HD utility. I use it on all new HD before I put HD's into service, and on my PC HD's every 6mo. In the FAQ's (link in my signature) is the procedure for replace the HD without loosing any data at the same time increasing capacity.

I would for sure remove a hd and make a copy of the boot tracks. Without this your 2200 will be dead if both HD's fail. Yes it take a little time but well worth it.

I notice that you are using Apple support. The newer 210/GuardianOS uses v3 SAMBA, which will give you longer file name support.

I personally like the speed of the larger 4 drive NAS's. The only problem with them is the fan noise, made for server rooms not office space. But I still use my 2200 on a reguarly. I think I have over 4yrs on my Seagate 300gig drives and still going strong.

[Mikey_Snap]

Just ordered a pair of Western Digital RE 250GB drives (WD2500SB) - should be here by Thursday - will run regular backups in the meantime.

I did a successful drive swap/upgrade on a Snap 1000 last year using a hardware based IDE drive duplicator. But that was a single drive without a mirror. How would the precedure differ with a 2200 with mirrored drives? Do I have to break the mirror first, transfer the image to the new primary drive, install both new drives and reinitiate the mirror? How would I tell which drive is the primary after I break the mirror?

Mike

[blue68f100]

All of your answers are in the FAQ's. You basicly use the 2200 to transfer the OS then you manually transfer the data from the old drive to new. You end up with a broken mirror, but you don't care, all of the data is on the new drive. Your original dirves are in tact but with a broken mirror, so you have a backup. The you install the second drive into the drive 2 position. Once it has the OS you have the snap re-create the mirror with all of the data tranfered to it.

The primary drive ID is 10000. The drive setup will show the new and broken drive ID's. so you will know which is which.

[Mikey_Snap]

I pulled a more detailed debug log and I don't see any evidence of an HD issue (although I have the two new WD (RE) drives sitting here ready to be installed). The PANIC errors always occur the same way (though different workstations are connected each time) and then the Snap Server locks up:

05/29/2008 3:08:56 38 D SMB | SMB : sesssetup Unicode pwd exceeded max len: 82, truncate to 24
05/29/2008 4:49:28 38 D SMB | SMB : close connection for JANINESOFFIC-PC - socket error=45
05/29/2008 11:56:40 39 F SYS | PANIC : sorecv 1
05/29/2008 11:56:40 39 F SYS | Call Stack :
$00152BD2
$0014FA36
$0014F67D
$001859A0
$00185A95
$0016EA6C
$0010413C
$001041C3
================================


"JANINESOFFIC-PC" is the only Vista PC on the network - but like I said above, it's always a different PC that gets the socket error as I look through the log.

[blue68f100]

05/29/2008 3:08:56 38 D SMB | SMB : sesssetup Unicode pwd exceeded max len: 82, truncate to 24

I do not see this vert often, but since the snap OS is 9 yrs old it does not like PW over 12chr.

Can you send be the complete log in html and I will take a look.

The Snap is very sensitive to timing issues. So if one drive is having a problem it can cause the server to go into to panic. I have had several through here that this has happened too. I have also seen system board failure too, but these are rare. A unit over heating due to fan failure can cause problems too. I had one that the fan would seam to randomly stop.

This is why I suggested using Spinrite to test the drives. This will either confirm or eliminate any drive problem.

If you don't have Spinrite you can break the array by removing 1 drive and see if the problem exist with a single drive. Then swap the drives and test the second one. If you do this you will have a broken raid array. The easiest way to rebuild the array is to force the mirror to individual config. Then use a utility of your choice to do a quick erase of the second drive. Then with all of the data on the pri (10000) drive you can create a new mirror. So if you decide to replace the HD you do not need to erase.

I like the spinrite. This way the mirror is not broken when the drives are reinstalled in there original position.

Any PC with Vista is not recommended. The SnapOS is not compatable with it. Some have beenable to work around most problems (bugs) but not all.

[Mikey_Snap]

The odd thing is I'm not using passwords at all. And I have not noticed the fan stopping.

Check your PM for the full debug log file.

[blue68f100]

This may all be related with OS corruption due to drive failure. AND/OF Vista is bring it down.

Vista and OSX 10.5 are not compatiable with the SNAP OS. OSX may work ok if you use the NFS and not the AFS (Samba v2). MS Vista not a chance.