Buffalo NAS-Central Forums

after a power cut my tera went into EM mode. i flashed firmware (as requested) to 2.00 and it came out of EM mode.

the four red status lights came on and flashed constantly along with the disk access lights flickering.

The web page had gone into Japanese mode.

after some 5-7 hours the disk access lights stopped but the constant flashing red lights remained.

i got talked through setting the web page to english and was able to see the disk config which said

'spaning mode' (the device should have been in RAID 5) but gave a total file size an 695 (the same volume size as before the problem)

since i could get no further and support just said delete the array and start again i went for the 2.10a firmware update. still no joy here, except the RAID mode now said RAID5.

I'm now running a 'performing RAID Scanning' but after this (will know in 5 hours) i guess my data is lost..

2 things and 1 question then,

1 looks like UPS's are absolutly required for this type of system,
2 since the device was new i had not finished backing up m data, and have lost all my Itunes files- its never to early to backup stuff


Question: is there anything else i can do? (i dont run linux) but could install it. - its there a way of mounting these disks in software RAID without the Tera and are there any utilities that may help me recover data?

1 a UPS is strongly recommended if you use the terastations.
2 i think we all learned it the hard way.

this posting is only a comment...it is no answer to the question.

_________________
LS1 (2.6 kernel, foonas svn1062, 750 GB, UBoot 1.2) & LS Pro (FreeLink/jtymod/GenLink, changes all the time)
Thx to all donators!

thanks mindbear. - now ive read a few pages looks like the next steps would be, get a ppc mac 4usb -pata caddies and try some soft raid stuff to try and get data back.

still many hours after starting the disk check, its still running and - the whole admin page is running really slow so i guess its busy.

Question: anyone know of a log file that the check may generate? (am usign 2.10a firmware so have a telnet client)

RAID Array 1

Status Performing RAID scanning. (62.0 % Complete)
Time Remaining 949.2 Minute
RAID mode RAID5
Disk Structure Disk 1,Disk 2,Disk 3,Disk 4
Total Capacity Not Obtained
File Format Not Obtained

ahh, due to my lack of linux know-how and general 'fidability'

i'd logged onto to telnet and tried to look at log files using vi i.e VI error.log
to exit vi i used 'Ctrl Z' (i bet that not the right thing to do)

looking over a process list with command 'ps aux' i found lots of the following

admin 4262 4.0 0.3 1896 440 ? R Oct19 45:34 vi error_log
admin 4674 4.0 0.3 1896 440 ? R Oct19 45:33 vi error_log
admin 4961 4.0 0.3 1852 440 ? R Oct19 45:33 vi ntp.conf

the 4.0 reprisents 4.0% CPU time. (doh) so i kill -9 the pids and hay presto the repair has speeded up.

AID Array Information
RAID Array 1

Status Performing RAID scanning. (63.4 % Complete)
Time Remaining 273.3 Minute

my mum always stop me to stop playing!

Reb.

ok the check finished, no change (does it make a log file?) before i go mad and start running some commands, and tracking down an old mac etc etc.

1. should i try the orinigal firmware? as when the device went into EM mode the support guys told me to flash to version 2 (which i did) then to 2.10a. but it strikes me that the 1.12 version of the firmware may be worth a short?

2. this is the displayed info from my 'disk management' screen, anyone care to guess the possible problem ?
RAID Array Information
RAID Array 1

Status Normal
RAID mode RAID5
Disk Structure Disk 1,Disk 2,Disk 3,Disk 4
Total Capacity Not Obtained
File Format Not Obtained

cheers all

A UPS is highly recommended with the TeraStations as any time that you get an unexpected shutdown on power up the RAID checking process is invoked - and that takes over 5 hours on a 1Gb teraStation. However before I got a UPS I had power cuts and never lost any data - although I had to put up with the 5 hours while the system did the RAID check on the disk when I restarted.

As to what the problem is I am not sure - the Disk Management screen is giving the sort of display one would expect if the disk appears to be unformatted. All logs tend to be held under /var/logs -but whether any of them will be of any use I am not sure. If you are logged in via linux then the basic flash process has worked fine - the problem is with your data partition (which the flash process does not touch).

Commands it might be worth trying under linux are:

which on my system comes back with

if you get that OK then you can try a command like

which comes back with (yours would probably vary according to the shares you have set up)

In terms of recovering, the normal recommendation would be to delete the array, reformat the disks and then rebuild the array. However I can understand you wanting to try and recover the data first. Also the delete/reformat/rebuild process is EXTREMELY time consuming.

thanks for the reply [b:ea8e2]itimpi[/b:ea8e2] and the firmware (you rock)

after the above post i power cycled the device, aswas just about to put old firmware back, incase the disk structure can only be read by it for some reason) and its started another raid check..
i'm going to allow it to finish. however, not sure if the checks are repairing anything and if so if itterations help. (prolly not is my guess)

i was hoping for a log file outputting what this raid check thing is doing or finding

df gives me the following....

Filesystem 1k-blocks Used Available Use% Mounted on
/dev/shm 15360 92 15268 1% /mnt/ram
root@HD-HTGL40F:~#

ls -l /mnt/array1
total 0
drwxrwxrwx 5 root root 37 Mar 31 2005 spool


what confuses my here (i have a Novell Netware and windows background) is that df doesnt show a filesystem for array1 but the ls command is showing /mnt/array1 ?? (which is a mount point right?)
now since a repair is active the volume, i would guess is offline - so the question is which bits do i not understand..

var/log/messages has some interesting bits in it (someone tell me to stop posting if they dont think this may be usful to others, but i'm clearly also fishing for help
--snip-- ([b:22fbc]hmm firmware/driver issue - guess not related)[/b:22fbc]
Oct 20 22:04:46 HD-HTGL40F kernel: PCI: 00:0e.2 PCI cache line size set incorrectly (0 bytes) by BIOS/FW, BUFFALO:it mus
t be zero for PPC8241/8245
--snip-- [b:22fbc](hda4 hdc4 hde4 hdg4 need to look up what this is and why their superblocks are pooly)[/b:22fbc]
Oct 20 22:04:47 HD-HTGL40F kernel: md: Autodetecting RAID arrays.
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 0000001e]
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 0000001c]
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 00000000]
Oct 20 22:04:47 HD-HTGL40F kernel: md: invalid raid superblock magic on hda4
Oct 20 22:04:47 HD-HTGL40F kernel: md: hda4 has invalid sb, not importing!
Oct 20 22:04:47 HD-HTGL40F kernel: md: could not import hda4!
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 0000001e]
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 0000001c]
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 00000000]
Oct 20 22:04:47 HD-HTGL40F kernel: md: invalid raid superblock magic on hdc4
Oct 20 22:04:47 HD-HTGL40F kernel: md: hdc4 has invalid sb, not importing!
Oct 20 22:04:47 HD-HTGL40F kernel: md: could not import hdc4!
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 0000001e]
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 0000001c]
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 00000000]
Oct 20 22:04:47 HD-HTGL40F kernel: md: invalid raid superblock magic on hde4
Oct 20 22:04:47 HD-HTGL40F kernel: md: hde4 has invalid sb, not importing!
Oct 20 22:04:47 HD-HTGL40F kernel: md: could not import hde4!
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 0000001e]
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 0000001c]
Oct 20 22:04:47 HD-HTGL40F kernel: [events: 00000000]
Oct 20 22:04:47 HD-HTGL40F kernel: md: invalid raid superblock magic on hdg4
Oct 20 22:04:47 HD-HTGL40F kernel: md: hdg4 has invalid sb, not importing!
Oct 20 22:04:47 HD-HTGL40F kernel: md: could not import hdg4!
Oct 20 22:04:47 HD-HTGL40F kernel: md: autorun ...
-snip---
Oct 20 22:04:47 HD-HTGL40F kernel: md: md1: raid array is not clean -- starting
background reconstruction
Oct 20 22:04:47 HD-HTGL40F kernel: md1: max total readahead window set to 768k
Oct 20 22:04:47 HD-HTGL40F kernel: md1: 3 data-disks, max readahead per data-dis
k: 256k
Oct 20 22:04:47 HD-HTGL40F kernel: raid5: device hdg3 operational as raid disk 3
Oct 20 22:04:47 HD-HTGL40F kernel: raid5: device hde3 operational as raid disk 2
Oct 20 22:04:47 HD-HTGL40F kernel: raid5: device hdc3 operational as raid disk 1
Oct 20 22:04:47 HD-HTGL40F kernel: raid5: device hda3 operational as raid disk 0
Oct 20 22:04:47 HD-HTGL40F kernel: raid5: allocated 4362kB for md1
Oct 20 22:04:47 HD-HTGL40F kernel: raid5: raid level 5 set md1 active with 4 out
of 4 devices, algorithm 2
Oct 20 22:04:47 HD-HTGL40F kernel: raid5: raid set md1 not clean; reconstructing
parity
Oct 20 22:04:47 HD-HTGL40F kernel: md: syncing RAID array md1
Oct 20 22:04:47 HD-HTGL40F kernel: md: minimum _guaranteed_ reconstruction speed
: 100 KB/sec/disc.
Oct 20 22:04:47 HD-HTGL40F kernel: md: using maximum available idle IO bandwith
(but not more than 100000 KB/sec) for reconstruction.
Oct 20 22:04:47 HD-HTGL40F kernel: md: using 124k window, over a total of 243481
024 blocks.
--snip----
looking at the following page the hda4 is nothing to do with the main data partition and i see these errors listed there.
[url=http://www.terastation.org/wiki/Harddisk_Layout:22fbc]http://www.terastation.org/wiki/Harddisk_Layout[/url:22fbc]

either the plot thickens, or i'm loosing it.

Reb

i cant downgrade my firmware, not sure why



the message is, i'm sure helpful in Japanese.

anybody get any ideas? i fooled around with the linkstation_version.txt file to ensure it names a newer firmware, doesnt make a difference.

Reb

hmm i don't now get the raid reconstructing messages in the /var/log/messages file and the device doesnt rebuild on startup, i do however get the following which is new...


Oct 21 08:42:47 HD-HTGL40F kernel: Starting XFS recovery on filesystem: md(9,1) (dev: 9/1)
Oct 21 08:42:47 HD-HTGL40F kernel: XFS: xlog_recover_process_data: bad clientid
Oct 21 08:42:47 HD-HTGL40F kernel: XFS: log mount/recovery failed
Oct 21 08:42:47 HD-HTGL40F kernel: XFS: log mount failed

odd because this is related to MD not MD1 (maybe the same thing for all i know) but this would seem to be releted to the ever present 'invalid raid superblock magic on hda4'

get in there!

after much playing arround i had a go with the XFS_repair option

root@HD-HTGL40F:/# xfs_repair /dev/md1
xfs_repair: warning - cannot get sector size from block device /dev/md1: Invalid argument
Phase 1 - find and verify superblock...
Phase 2 - using internal log
- zero log...
ERROR: The filesystem has valuable metadata changes in a log which needs to
be replayed. Mount the filesystem to replay the log, and unmount it before
re-running xfs_repair. If you are unable to mount the filesystem, then use
the -L option to destroy the log and attempt a repair.
Note that destroying the log may cause corruption -- please attempt a mount
of the filesystem before doing this.

since the file system could not me mounted i went for it

root@HD-HTGL40F:/# xfs_repair -L /dev/md1 It worked!!!!

now this replaced my 'share' directory (which has some data in that i need -backing off to pc now)

but i had lots of other shared folders which i cannot see.

Question: where will this data be held? it looks like the array contains the same amount of data as it should so i expect its all still there - somewhere, but since the device has been flashed many times the user and share information is no more.

maybe this is some samba config? help!

feel like im getting somewhere.

reb

on the console i can search for and find the files that i can't see via connecting to a share so it is there (excepting that clearing the transction log may have lost/corrupted some data)

just need to be able to copy it off now. any help allowing a networked device (windows) connect to this data that appears not to be shared would be appreciated.

Reb

sorry for spamming the forum.

looks like i've got it all back

to get the shares again just recreate them via the web page using the correct case and hey presto the files are back!!!

lost and found is empty so maybe its all fixed!!

/me starts many many backups!!

Phew! In addition to the backups, get a UPS!

_________________
HG · Debian squeeze
Linux Kernel 2.6/PPC (and more!) for Linkstation/KuroBox/HG/HS/Terastation from http://hvkls.dyndns.org

just to round up i think the raid check (after running 3 times) may have been a bit of a red herring, although it did autorun so perhaps it did fix something?

the system seems to run an XFS_repair on startup to try and recover the XFS volume but it omits the -L switch which clears the transaction log.
It would seem my transaction log was corrupt so the repair could not continue.

running the with the -L command allowed the process to run and complete. the only file it complained about was
'entry "backuplog3_200610190600.txt" at block 0 offset 456 in directory inode 701617076 references free inode 701617025
clearing inode number in entry at offset 456...'

which was the RSYNC process running when the power failure happend, and sure enough that file now no longer exists.

Thank you all for your help in this matter. and big thanks to itimpi without who i'd have never had telnet access to get into the system

andre - in process of purchasing UPS as we speak!

Regards,

Reb