Snap 1000 Error

[microsvc]

I have been getting the following errors on my Snap server 1000 which has been running flawlessly for a year. It has a 500gb HDD which I upgraded to with help of previous posts here.

fsck fatal error = 8
Failed to allocate 3798204 bytes for i-node cache buffers.

Can anyone advise what can cause this and how I may resolve it?

OS 4.0.860
HW 3.1.2
Bios 2.1.366

Thanks!
~bOb

[Terry Kennedy]

Originally Posted by microsvc

I have been getting the following errors on my Snap server 1000 which has been running flawlessly for a year. It has a 500gb HDD which I upgraded to with help of previous posts here.

fsck fatal error = 8
Failed to allocate 3798204 bytes for i-node cache buffers.

Can anyone advise what can cause this and how I may resolve it?

On the older Snap models, is is possible to install disk(s) which are too large to be checked, due to the small amount of RAM in those models. Normally they will fall back to a swap-based fsck (which takes appoximately forever to complete), but it sounds like you have a case where even that doesn't work.

About all I could suggest is to disable any extensions you may have loaded, like the JVM. It might be possible to free up a small amount of memory with debug commands (a 1000 doesn't need RAID buffers, for exmple), but it has been years since I looked at this. Perhaps someone else with more current info can advise.

[bitor]

The JVM doesn't apply to the Snap1000. Not enought memory on board for it. Try running HH Regenerator to check for HDD errors. You can also use Spinrite too for errors on HDD.
Report back. There may be another way around the fsck if the checking of the HDD for errors doesn't work.

Originally Posted by microsvc

I have been getting the following errors on my Snap server 1000 which has been running flawlessly for a year. It has a 500gb HDD which I upgraded to with help of previous posts here.

fsck fatal error = 8
Failed to allocate 3798204 bytes for i-node cache buffers.

Can anyone advise what can cause this and how I may resolve it?

OS 4.0.860
HW 3.1.2
Bios 2.1.366

Thanks!
~bOb

[microsvc]

Update: I removed the 500gB drive and put in the old 30gB drive the server had originally. Its working fine on the 30gB. I have been researching these forums and found that many cant get 500gB to work on a 1000, but I did for a bit. Odd.

I also read that 320gB works in all cases, so I'm going to take that route. I'm going to run Spinrite on this drive and report the results, but I think the drive is fine. If that's the case, how can I recover the data from this 500gB drive?

What's the difference between a 1000 and a 1100?

Would the 500gB work on an 1100?

Thanks for all the responses to my three questions!
~bOb

[bitor]

My thoughts on the 30GB working and not the 500GB working. One thing that comes to my mind would be a power supply issue. A way to test this is to take a PC power supply that's around the house and hook up ONLY the hard drive. Then power on the Snap1000 and see if you still have the same problem.

Another thought would be the fiscal lack of RAM in the 1000=32MB; 1100=64MB.
There is a way via com1 on the Snap4000 to turn fsck off, BUT I cannot remember the command AND I don't know if they removed this with your SnapOS.
I'll have to hook up the com1 and load the same version SnapOS you have and see if it is present like it was on the earlier version I saw via com1 on Snap4000. On top of this, if it's applicable for the Snap1000 is another consideration.

I had this happen to me on my 250GB Snap1000. How I got around it was to upgrade the snapos from v3.4.805 to your newest version and that fixed it.
This could have been caused by a file os corruption and reloading it could have corrected it OR the upgrade could have. Either way , it's an option.

Snap1100 also has an updated BIOS and 64MB other then that they are basically the same unit. The 500GB should work in either or the 1000 OR 1100. But I would keep the Snap1100 over the Snap1000 just based on the extra RAM 64MB.


I would do this on Snap1000.

1.) Test the hard drive. checks ok then
2.) Power only the hard drive with and external PC power supply if no change
3.) Try to reinstall via Assist your SnapOS again.

Recover question should be addressed after all else fails.

Note on "2.)" This is just to test and not to be run contently without some sort of cooling fan on the unit otherwise the unit may over heat.

Originally Posted by microsvc

Update: I removed the 500gB drive and put in the old 30gB drive the server had originally. Its working fine on the 30gB. I have been researching these forums and found that many cant get 500gB to work on a 1000, but I did for a bit. Odd.

I also read that 320gB works in all cases, so I'm going to take that route. I'm going to run Spinrite on this drive and report the results, but I think the drive is fine. If that's the case, how can I recover the data from this 500gB drive?

What's the difference between a 1000 and a 1100?

Would the 500gB work on an 1100?

Thanks for all the responses to my three questions!
~bOb

[microsvc]

Thanks for your assistance. See below.

1.) Test the hard drive. checks ok then
I ran DD Regenerator on the drive and it found no errors

2.) Power only the hard drive with and external PC power supply if no change
I powered the drive externally and there was no change in functionality.

3.) Try to reinstall via Assist your Snap OS again.
I ran the Assist program and chose the SUP file and it said successful.

Nothing has resolved the problem. If you think I have run out of options, then:
Recover question should be addressed after all else fails.

Can I shut off FSCK somehow so this drive will mount?

Anything else?

~Bob

[bitor]

Ok, great work. Give me sometime to setup my Snap4000 and install the newest OS and check the command via com1.

I'll report back in a few days. Thanks

Originally Posted by microsvc

Thanks for your assistance. See below.

1.) Test the hard drive. checks ok then
I ran DD Regenerator on the drive and it found no errors

2.) Power only the hard drive with and external PC power supply if no change
I powered the drive externally and there was no change in functionality.

3.) Try to reinstall via Assist your Snap OS again.
I ran the Assist program and chose the SUP file and it said successful.

Nothing has resolved the problem. If you think I have run out of options, then:
Recover question should be addressed after all else fails.

Can I shut off FSCK somehow so this drive will mount?

Anything else?

~Bob

[bitor]

Ok, try this before I start to rearrange my Snap4000.

At debug command type the following.

type:
config device fix disable
then type:
config device force disable
then type:
reboot

After rebooting and if all goes well and it's still not mounting then issue the command at the debug line
type:
config device mount
(You may have to reboot if you do not see your folders.)

Report back, Thanks

Also forgot to tell you that the Snap1100 has a faster CPU, but not much better then the Snap1000.

[microsvc]

type:
config device fix disable
I get
"Partial command entered, one of:
config devices automount enable|disable (Auto-Mounting of devices at boot)
config (individual/mirror/span/raid/xtra (=spare)) [dev,...]
format dev [label ["comment"]] [/reinit [/nocore]] (format an unmounted device)
fsck dev [[/fix /fixfatal /altsb]] (check device's filesystem)
small SizeInMB dev0 ... devN
info (same as "info devices")
mount [dev] request to mount one or all devices
predict (individual/mirror/span/raid) [dev,...]
resync dev
testparity dev
unmount [force/dev] request to unmount one or all devices
Command returned with console error -2."

then type:
config device force disable
I get
"config device force disable
Partial command entered, one of:
config devices automount enable|disable (Auto-Mounting of devices at boot)
config (individual/mirror/span/raid/xtra (=spare)) [dev,...]
format dev [label ["comment"]] [/reinit [/nocore]] (format an unmounted device)
fsck dev [[/fix /fixfatal /altsb]] (check device's filesystem)
small SizeInMB dev0 ... devN
info (same as "info devices")
mount [dev] request to mount one or all devices
predict (individual/mirror/span/raid) [dev,...]
resync dev
testparity dev
unmount [force/dev] request to unmount one or all devices

Command returned with console error -2."

then type:
reboot

After rebooting and if all goes well and it's still not mounting then issue the command at the debug line
type:
config device mount
I get "Unable to rescan all devices because dduQuietAllDevices() was not previously called
------------
Command returned with console error 1."

(You may have to reboot if you do not see your folders.)

Sorry, this didn't work

[bitor]

microsvc, I'm going to have to load the newest os and go in to com1 to see if the newest has a command to disable fsck. Those commands are from os 3.4.7 via Snap4000. If I find the command in the newest version then it doesn't apply to the Snap1000. Most if not all the commands cover all the SnapOS units. But I've seen commands come and go in diff. SnapOS versions.

I'll do this today and report back.

[bitor]

microsvc,

I've looked at my Snap4000 via com1 and it does have these fsck commands in 4.0.855. So, most likely it would have it in 4.0.860. So these commands are not working with the Snap1000.

Have you tried to let fsck run all the way throw to see if it will correct again?

Try this at the debug command line.
conf dev automount disable
then,
reboot
then debug command line again,
conf dev mount all

Now check to see if you can see the folders. If this doesn't work then recovery may be an option for you. See below.

You can try to get a hold of user "snap-tech, snap-tech-dude" here. They are both the same person. He does recovery for SnapOS and is reasonably priced last time I heard. Either way on the recovery will not be cheap.

Douglas aka snap-tech,snap-tech-dude
email:
dfullersnap@yahoo.com
or
douglas@frontlinedatarecovery.com
Cell #:425-244-0302(this might be an old number)



Good luck and keep us posted.

[microsvc]

The first command automount disable succeeded.

The mount all gave me this:
06/09/2010 17:13:00 Command: conf dev mount all
Unable to rescan all devices because dduQuietAllDevices() was not previously called
------------
Command returned with console error 1.

Ideas?

[bitor]

Try this at the debug command line again and see if you get the
dduQuietAllDevices without the reboot this time.

conf dev automount disable
then,
conf dev mount
then,
conf dev automount enable

check your folders to see if they are present.

Originally Posted by microsvc

The first command automount disable succeeded.

The mount all gave me this:
06/09/2010 17:13:00 Command: conf dev mount all
Unable to rescan all devices because dduQuietAllDevices() was not previously called
------------
Command returned with console error 1.

Ideas?

[microsvc]

Hi, I get the dduQuietAllDevices after I enter conf dev mount.

Do you know what the format is of the drive? Can I use Linux to read the drive?

Thanks for all your help!

[bitor]

There IS a recovery program you can use to access the hard drive called UFS Explorer. I have used UFS Explorer and it worked for me. The SnapOS is a modified BSD OS having a UFS. You can try Linux, but I don't think it will work. Check out the link below which will give you some knowledge and put you on a good path.

Info on Unix File System(UFS)
http://en.wikipedia.org/wiki/Unix_File_System

Can you unmount and then mount the drive without getting the "dduQuietAllDevices" ?

conf dev unmount
conf dev mount

Originally Posted by microsvc

Hi, I get the dduQuietAllDevices after I enter conf dev mount.

Do you know what the format is of the drive? Can I use Linux to read the drive?

Thanks for all your help!

[DCC1234]

Can the 500 gig drive be formatted to a lower capacity-- say 400gig- to eliminate the out of memory error on the 1000. How to format for lower capacity?

[bitor]

Yes, you can format a HHD via debug command to any size you want below the max. physical capacity of the HHD. It's been a while since I've issued this command.
From the debug command line

conf dev small SizeInMB 10000

SizeInMB=put what size you want the drive to be here in Megabytes.
10000= your hard drive which in this case is a Snap1000/1100 model.
It will take sometime for it to complete the task. Go in the WebGUI to the disk area and check to see if it is resizing the HHD. If it isn't then reboot and see if it is. Like I said, I've issued these commands and they have worked , but it's been a long time ago.

Make sure you back up your data on your HHD before issuing this command to be safe.
Let me know how it goes.

Originally Posted by DCC1234

Can the 500 gig drive be formatted to a lower capacity-- say 400gig- to eliminate the out of memory error on the 1000. How to format for lower capacity?

[DCC1234]

Thank you. But this command seems to cause it format again at full capacity.
Any other suggestions?

CONFIG DEVICES SMALL 400 10000

[bitor]

It works you have it set to format at 400MBs. which you wanted 400GB
So, 400GB = 409600MB

Try this command and no need to reboot.

conf dev small 400000 10000

I'll recheck on my server to make sure, but like I said it's been a while, but I do remember it working for me when I did use it.

Originally Posted by DCC1234

Thank you. But this command seems to cause it format again at full capacity.
Any other suggestions?

CONFIG DEVICES SMALL 400 10000

[DCC1234]

All I can say is duh!! Will try again tomorrow and report back....

[DCC1234]

typed that exact command and it is formatting now, unfortunately--- to the full capacity. oh well. I'll like for a utility from the manufacturer (WD). or just give up on this!

[bitor]

Hold on, let me go to my server and check this out. I've used it in the past and it's worked for me.

brb

Originally Posted by DCC1234

typed that exact command and it is formatting now, unfortunately--- to the full capacity. oh well. I'll like for a utility from the manufacturer (WD). or just give up on this!

[bitor]

I just tried the command on my Snap1000 and it worked.

Try this command and no need to reboot.

conf dev small 40000 10000


This should give you 40GB

Also, go to debug and type:

conf dev info

and copy and paste what it says here.

Are you using a PATA/IDE drive or SATA drive with an adapter?

[DCC1234]

500 GIG AAKB . Re-imaging SNAP O/S onto the drive now. I'll send debug info when done.
At this point-- this may be an academic exercise-- but I appreciate you hanging with me.

[DCC1234]

command shows 14gig after I imaged o/s. But its a 500 gig drive that gets hosed every time I try something... Here is conf dev info:


Logical Device: 10006 Position: 0 JBOD Size (KB): 32296 Free (KB): 24000 Private Mounted
Label:Private Contains system files only
Unique Id: 0x544508362E4EF06C 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): 14668288 Fixed

Logical Device: 10000 Position: 1 JBOD Size (KB): 14524624 Free (KB): 0 Public Unmounted
Labelrive1 Single disk
Unique Id: 0x544508362E4EF06C Mount: /0 Index: 0 Order: 255
Partition: 10000 Physical: 10007 FS Size (KB): 14524624 Starting Blk: 17830 Public
Physical: 10007 Drive Slot: 0 IDE Size (KB): 14668288 Fixed