Snap M4100 failde after battery replacement

[Mauro]

Hi All,

before I would like to excuse if my English isn't fluent: I'll try to explain better than I do.

A friend of mine has a Snap M4100 that start with replace CMOS battery message: the sw company who assist him replace the battery with a new one but, after that, the server become unreacheable on the net ( both Assist and web interface ) and it seems also that the disk raid is not working/loaded. He ask me for an help - I work as IT professional for a private company - but I've never managed this kind of server. Technician of the sw company told him that could be a raid problem, but I don't think that a battery replacement could disrupt a raid config or scratch a flashrom based OS. Unfortunately I've not much technical data from the server ( OS release ecc. ) except that it's - I think, after reading some related documentation - a raid 5 config and that hd set is 3x60Gb and 1x80Gb.....and this appear to me a very strange kind of configuration.

Before going to operate on the server, could anyone of you professional give me some advice or ideas ?

Many thanks to all.
Cheers
Mauro

[achapman]

Presumably replacing the battery would mean that the the unit would forget its configuration information; it would be equivalent to a factory reset. This would make it lose its IP address, but its data should still be intact as the unit should recognise the disks it contains as having a valid configuration.

[blue68f100]

If this is the case you may need to use assist to find the unit and do the initial setup. If I recall the earlier units def to 0.0.0.0 . I have seen assist have problem when run under XP. Works best with Win98. If you use XP set the compatability and turn temp turn off the personal firewall.

[Terry Kennedy]

Originally Posted by Mauro

I think, after reading some related documentation - a raid 5 config and that hd set is 3x60Gb and 1x80Gb.....and this appear to me a very strange kind of configuration.

Others have pointed out that the config may have been reset when the battery was changed.

Regarding the 1 drive of a different size, that was probably from a warranty replacement. As the old drives were obsoleted, Snap shipped larger replacements. For example, I received a 75GB replacement for a failed 60GB drive. The SnapOS default is to initialize a replacement drive to the same size as the remaining drives (assuming it is big enough). The extra space is unused unless you go into the debug menu and re-init the drive to use all available space.

[Mauro]

OK, thanks to all for reply. I've had tried with Assist from OS X Leopard to find the server, but no luck: my mind is try to pull out all disk from server - simply detach the flat cable from mobo - and give it a try to see if could be a hw issue. Anyway, could be possible retrive data if server if definitive loss ? I've read that attaching to a Win machine it's not good because of a loss of data, but running third party app from a boot cd it's possible ? I would like to try with RStudio for recovering data and SpinRite for a disk check.

Thanks.
Mauro

[blue68f100]

Try the PC Version of Assist, I have never tried the Mac Version. Do a hard reset using the 4 blink to reset every thing and try again. Remember this is 10yr old hardware and technology. It works best using OS that were in that era. There is actually 2 different pc's version out, one for v3 and another for v4. The latest works with both.

Mark you slot/drive location if you remove the HD's. Normally a larger HD in slot 1 causes problems. Also depending on how the software was upgraded and hd's formatted under the older file structure also can cause problems.

I would locate another 4100 and install your drives to extract the data. If yours do not start up. If it does and will not boot off of your array you may have a bad power supply.

But removing the drives and seeing if the server shows up is a good idea.

[Mauro]

Hi guys,

thank to blue68 for suggestion. Here again with good/bad news.

Good:
after detaching hd from mobo, I was able to reach the server both with Assist and web browsing.

Bad:
with hd detached, I can't logon to it because my friend don't know the admin password
( I know that it could be solved with hard reset, but I would like to leave this as ultimate
beach because I've read more other user post where they are unable to login with standards
after that )
all volumes container are signed with a red cross on the folder symbol with "unavailable disk.."
description on the right ( obviously this is normal )
reattached hd to mobo, server become again unreacheable both with Assist an web browsing

Now...from my point of view, raid config is gone, but I don't know few this os, so it's only
a suspect.

I'm now running Spinrite on first hd ( 80 Gb ) and a couple of thing made me crazy:
Spinrite is running from boot cd, the drive is attached externally through USB, and the
sw say me that the drive is empty ( ARGH !!! ) and that is an Add-on controller....any suggestion ?

Thank for the support
Cheers
Mauro

[blue68f100]

Check the log and see if it was running a check on the HD's. If so it will not mount the raid till it has finish. I have seen this take several days if the HD's are over 80% full. Every time you stop it it starts over.

After SpinRite let it run and check the log for progress.

You can only reset the password without blowing the other settings out. Need to see how many blinks are needed for this. I know it's not 4=full reset all.

[Mauro]

Thank blue68.

Do you think some sw like RStudio or other similar could help to pull out data from the raid ?

I don't really want put more time on it, also because - my friend apart - I will not receive nothing as payment

Thanks.
Mauro

[Mauro]

Oh, sorry, forgot to mention....Spinrite is running on level 2 at this moment. It is the Recover unreadable data mode: perhaps, could some other level a better way ?

Cheers
Mauro

[Mauro]

An image is better than hundred words.

This is taken after rebooting server and trying to reach by web GUI: Spinrite log didn't show nothing wrong on the first disk....no damaged or recovered sector.

No more ideas on why raid group is offline. What about total reset ? Could be helpful ?

Last....Assist was able to discover the server and figured it green ( configured ).

Thanks
Mauro

[blue68f100]

Like I said if it's checking the disks it will show off line. Doing a reset will not cause any day lost. Personally after the spinrite check I would install all of the HD's and restart it. After about 15 min, then look at the log. The log will tell you whats going on.

Are any of the LED lights lite up on the front panel?
If a HD is reported as bad it will show up there.

[Mauro]

Thanks blue68. I think there was a misunderstood: I'm checking disk once at time, attached externally USB to my pc booted from Spinrite cd. This for checking if one of the hd had error on itself: I forgot to mention that I've got disks unavailable AFTER reattaching them to snap mobo and access the server from web GUI.

One question: if I copy sector by sector all disk to other with more capacity ( 80 Gb instead 60 of the original mounted ) and put them on the server, could be it alive again or maybe disk still offline ?

Again, thanks for precious support, and sorry to bother with stupid question but I'm really newbie on this kind of product.

Mauro

[blue68f100]

Doing a Raw copy works, but it will report an error due to size difference. Spinrite could correct the end sector problem.

You might want to install a couple of new HD in the 4100 and see if it see's them correctly. You may have a hardware problem if it does not. If you have the sup file, you may try do a reinstall of the OS. It's possiable that something had a hickup.

At this point and time I would do a full reset. Then test new HDs. If good, reinstall the orig HD's. Having a larger HD in slot 1 has caused a lot of problems. The factory had mfg hd utilities that allowed them to change the size and make them worked. The Snap uses drive 1 capacity to setup the array. If this is larger it messes it up. There has also been some HD that have problem setting the buss speed corectly, ATA100. 133 is not supported on the older units.

If it is seeing the HD's,NO Errors on front panel, leave it on for a couple of days. Every time you power off it will run chkdsk utility, which can take for ever if the HD is over 80-90% full.

[Mauro]

Full reset with newly hd installed....sounds good but.....what about if I do this with older hd on ? Maybe data loss ?

Thanks
Mauro

[blue68f100]

Doing a full reset does not damage any data, this is preserved. It only impacts admin pw, network settings and 2 other minor settings.

I have done this many times on SnapOS units, 1x00, 2x00 units and 4100's.

[Mauro]

OK...finally I entered on server os and run debug: here results

Logical Device: 10006 Position: 0 JBOD Size (KB): 32296 Free (KB): 19952 Private Mounted
Label:Private Contains system files only
Unique Id: 0x4321BC1D49A57614 Mount: /priv Index: 12 Order: 0
Partition: 10006 Physical: 10007 FS Size (KB): 32768 Starting Blk: 515 Private
Physical: 10007 Drive Slot: 0 IDE Size (KB): 78150656 Fixed

Logical Device: 1000E Position: 0 JBOD Size (KB): 32296 Free (KB): 19952 Private Mounted
Label:Private Contains system files only
Unique Id: 0x137BDEDB4F8217F4 Mount: /pri2 Index: 13 Order: 1
Partition: 1000E Physical: 1000F FS Size (KB): 32768 Starting Blk: 515 Private
Physical: 1000F Drive Slot: 1 IDE Size (KB): 58615296 Fixed

Logical Device: 10008 Position: 1 ORPHAN Size (KB): 58154080 Free (KB): 0 Public Unmounted
Labelrive2 Orphan from BACKUP - RAID5
Unique Id: 0x7484411026802832
Partition: 10008 Physical: 1000F ORPHAN Size (KB): 58154080 Starting Blk: 57524 Public
Physical: 1000F Drive Slot: 1 IDE Size (KB): 58615296 Fixed

Logical Device: 10018 Position: 2 ORPHAN Size (KB): 58154080 Free (KB): 0 Public Unmounted
Labelrive4 Orphan from BACKUP - RAID5
Unique Id: 0x7484411026802832
Partition: 10018 Physical: 1001F ORPHAN Size (KB): 58154080 Starting Blk: 57524 Public
Physical: 1001F Drive Slot: 3 IDE Size (KB): 58615296 Fixed

Logical Device: 60000 Position: 3 RAID_BROKEN Size (KB): 174462240 Free (KB): 0 Public Unmounted
Label:RAID5 Large data protection disk
Unique Id: 0x7484411026802832
Partition: 10000 Physical: 10007 R 60000 Size (KB): 58154080 Starting Blk: 69734 Public
Physical: 10007 Drive Slot: 0 IDE Size (KB): 78150656 Fixed
Partition: 10010 Physical: 10017 R 60000 Size (KB): 58154080 Starting Blk: 57524 Public
Physical: 10017 Drive Slot: 2 IDE Size (KB): 58615296 Fixed

any way to have the two orphaned drive recovered ? I know that on a raid 5 with 4 disk it's impossible, but if someone has good suggestion will be welcome.

I was thinking to connect the server crossed to my laptop/desktop and see if I can recover data with some sw tool ( RStudio, UFS Explorer or Raid data recovery ). Could be a right way ?

Thanks.
Mauro

[blue68f100]

The only way to recover from 2 Orphans is by a recovery service. They can strip the data off the failed drives. Snaptech does recovery service for these old units. He use to work for Quantum HD before Snap Appliance bought them. He can do it just search for his post for contact info.

[Mauro]

OK...it was what I expected to hear.

At this point - except Douglas as ultimate beach - I'm trying to rebuild the array imaging each drive with DDRescue cmd on Linux live system and then using them with Raid Reconstructor to see how much fles could be recovered and - more important - if it's possible. But...because two of the drives return I/O error due to bad sectors, what I want to know is what could be happen attaching them USB to a Win 7 machine for using Seagate Tools to check if they are repaired.

I've read that attaching a Snap hd to Win system cause loss of all data: in theory, USB work less in background than a system disk because of no request until he his invoked to act. It's really possible to loss data simply at the "new hw found" option by Windows ?

Again, many thank for support.
Cheers
Mauro

[blue68f100]

It's really possible to loss data simply at the "new hw found" option by Windows ?

My experience YES

MS does not follow users request. Like ReStart now or later. It keeps asking. If you do not respond it will restart, loosing all of your open data.

[achapman]

I'd also agree. Windows OS like to write a disk ID to the disk which then makes it unusable under the original OS. While sometimes you are warned I'm not sure it is always the case. I wouldn't risk it myself, having been caught out by this several times.

[Mauro]

Well, thaks to both for experience. I'm now try to img disk with DD_Rhelp: it more powerful and speedy with hd that have bad sectors.

I will let you know results when all steps will be done: if I've raid reconstructed, this could be helpful for others with same issue.

All my best
Mauro