Lacie 5Big2 Failed After Firmware Update

[fvdw]

rebooted itself... :disapprove

there is no script in the standalone kernel that guards the temperature.
The only thing I can think of is a power cut, that will cause a reboot.
Even a fatal error of the kernel won't initiate a reboot

ps if this happened while you were doing the Lacie upgrade it could be a reason why it failed
I guess it rebooted somewhere in that process, but if that reboot was initiated by something else the the firmware upgrade it can cause big issues...

Check you power supply would be my advice

[Mijzelf]

makes the raid

The raid array only have to be created once. (Which already is done) Next time it has to be assembled:

Code: Select all

mdadm --assemble /dev/md0 /dev/sd[abcde]2

where the sequence of abcde isn't important anymore.

So the whole seqence is:
Upload mdadm to the box
execute:

Code: Select all

mdadm --assemble /dev/md0 /dev/sd[abcde]2
mkdir /mp2
mount -o ro /dev/md0 /mp2
tcpsvd -vE 0.0.0.0 21 busybox ftpd /mp2

[Mijzelf]

I'm only getting about 26Mb per second

I wouldn't be surprised if you can double that by simply starting a 2nd ftp daemon on another port, and let 2 clients suck 2 different directories.
As long as the cpu load (as seen in top) stays below, let's say, 80%, there should be room for improvement.

But I like the idea o a script to certify the data.

OK, I'll write something.

[fvdw]

makes the raid

The raid array only have to be created once. (Which already is done) Next time it has to be assembled:

Code: Select all

mdadm --assemble /dev/md0 /dev/sd[abcde]2

where the sequence of abcde isn't important anymore.

So the whole seqence is:
Upload mdadm to the box
execute:

Code: Select all

mdadm --assemble /dev/md0 /dev/sd[abcde]2
mkdir /mp2
mount -o ro /dev/md0 /mp2
tcpsvd -vE 0.0.0.0 21 busybox ftpd /mp2

your forget to upload and run udevstart

IMPORTANT

Remember this is a kernel running from RAM, everything you uploaded to RAM and settings you changed in the root file system is lost when rebooting

[Mijzelf]

your forget to upload and run udevstart

Actually I didn't. When assembling a raid array the role is read from the headers, and so the sequence as given to mdadm isn't important anymore. Fortunately. You can get the disks out and shuffle them, and still the array will assemble without problems.

I wrote a script, as attached.

/Edit: I was not yet ready, but hit submit, somehow.

Upload mdadm, udevstart and the tgz file. Run udevstart, extract the tgz file. The script generate_checksums.sh will daemonize and put 7 files in the current directory, abcd.txt. abce.txt, ... and stdout.txt and stderr.txt.
So you can create a directory, and run it from there:

Code: Select all

mkdir /testdir
cd /testdir
/path/to/generate_checksums.sh

I have no idea how big the files can get. Especially stderr.txt. It is possible that some incarnation will give file system errors, in which case the file could be huge. I *think* it will fit in the 512MB memory, but maybe you'd better mount an usbstick (I suppose the kernel supports FAT), and run it from there. It has also the advantage that you won't loose already completed incarnations on a spontaneous reboot.

In case of a spontaneous reboot the script has to be changed a bit, to omit the already completed incarnations.

The script daemonizes, which means you can close the telnet session. The script should just continue working.

Oh, and on start the script expects the raid array to be not assembled.

/Edit2: Don't start this before you're done copying.
Check if md5sum is available, before starting.

[unlucky1]

Okay, after about half a day of moving files (doing well) ftp would fail but I could get it going again. At some point it completely failed and booted me out of ftpd so if I tried to go in again, it would hang. So I started checking and found that it opened a gazillion ftpds (gazillion being a number too high for me to count or a non-arabic number that my keyboard can't show). So, how to kill all these ftpds without rebooting (showing now just 1 page of a gazillion):

Code: Select all

    1 root       0:01 /bin/sh
    2 root       0:00 [kthreadd]
    3 root       0:01 [ksoftirqd/0]
    6 root       0:00 [khelper]
    7 root       0:00 [kworker/u:1]
  204 root       0:00 [sync_supers]
  206 root       0:00 [bdi-default]
  207 root       0:00 [kblockd]
  212 root       0:00 [ata_sff]
  223 root       0:00 [khubd]
  230 root       0:00 [md]
  250 root       0:00 [rpciod]
  258 root       6:05 [kswapd0]
  306 root       0:00 [fsnotify_mark]
  320 root       0:00 [unionfs_siod]
  321 root       0:00 [nfsiod]
  339 root       0:00 [jfsIO]
  340 root       0:00 [jfsCommit]
  341 root       0:00 [jfsSync]
  348 root       0:00 [xfs_mru_cache]
  349 root       0:00 [xfslogd]
  350 root       0:00 [xfsdatad]
  351 root       0:00 [xfsconvertd]
  352 root       0:00 [crypto]
  465 root       0:00 [scsi_eh_0]
  468 root       0:00 [scsi_eh_1]
  471 root       0:00 [kworker/u:2]
  478 root       0:00 [smflush]
  487 root       0:00 [orion_spi]
  492 root       0:00 [mtdblock0]
  497 root       0:00 [mtdblock1]
  562 root       0:00 [kmpathd]
  563 root       0:00 [kmpath_handlerd]
  617 root       0:00 [mv_crypto]
  730 root       0:00 /sbin/telnetd
  734 root       0:00 /sbin/buttons-nwsp2
  735 root       0:00 -sh
  743 root       0:00 -sh
  827 root       0:00 [md0_raid5]
  842 root       0:00 [xfsbufd/md0]
  849 root       0:00 [ftpd]
  850 root       0:00 [ftpd]
  851 root       0:00 [ftpd]
  852 root       0:00 [ftpd]
  853 root       0:00 [ftpd]
  854 root       0:00 [ftpd]
  856 root       0:00 [ftpd]
  857 root       0:00 [ftpd]
  858 root       0:00 [ftpd]
  859 root       0:00 [ftpd]
  860 root       0:00 [ftpd]
  861 root       0:00 [ftpd]
  862 root       0:00 [ftpd]
  863 root       0:00 [ftpd]
  864 root       0:00 [ftpd]
  865 root       0:00 [ftpd]
  866 root       0:00 [ftpd]
  867 root       0:00 [ftpd]
  868 root       0:00 [ftpd]
  869 root       0:00 [ftpd]
  870 root       0:00 [ftpd]
  871 root       0:00 [ftpd]
  872 root       0:00 [ftpd]
  873 root       0:00 [ftpd]
  874 root       0:00 [ftpd]
  875 root       0:00 [ftpd]
  876 root       0:00 [ftpd]
  877 root       0:00 [ftpd]
  878 root       0:00 [ftpd]
  879 root       0:00 [ftpd]


[fvdw]

you could try

Code: Select all

killall -9 ftpd

or als this should do it (if awk is present in the standoanlone kernel, I forgot)

Code: Select all

$ kill -9 `ps -ef | grep ftpd | grep -v grep | awk '{print $2}'`

Does this happens after a reboot when stating ftpd for firs time ?

[unlucky1]

I tried killall and kill in it's various circumstances. I tried your kill grep with no success:

Code: Select all

root@(none):/ # kill -9 `ps -ef | grep ftpd | grep -v grep | awk '{print $2}'`
-sh: can't fork
sh: you need to specify whom to kill

It happens as ftpd seems to run. When I started it it was just a few instances. It appears to have grown quite a bit over time. At some point the OS couldn't handle that many as it wouldn't let me start new ones.

Here's the tail-end:

Code: Select all

 8863 root       0:00 [ftpd]
 8864 root       0:00 [ftpd]
 8865 root       0:00 [ftpd]
 8866 root       0:00 [ftpd]
 8867 root       0:00 [ftpd]
 8868 root       0:00 [ftpd]
 8869 root       0:00 [ftpd]
 8870 root       0:00 [ftpd]
 8880 root       0:39 [kworker/0:2]
 8882 root       0:00 [ftpd]
 8947 root       0:15 [kworker/0:0]
 9058 root       0:00 -sh
 9071 root       0:00 [ps]
 9072 root       0:00 [grep]
 9073 root       0:00 [grep]
 9077 root       0:01 ps -a


[unlucky1]

Eeep! I tried your script again with ftpd in quotes and it busted me outta shell and now it's in fork hell (where apparently there is no spoon). So even if I start a new terminal, I'm still there:

Code: Select all

(none) login: root
Password:
-sh: can't fork
~ #
~ # ps -e
-sh: can't fork
~ #
~ #
~ #


[fvdw]

well it seem to have killed everything :whistle thats what you get when doing something you don't know, the quotes are there for a reasons :disapprove

you will need to reboot :mrgreen:

that the number of ftpd's grows in time, might be because there is a problem with the disk on the part where it wants to right, but I am no sure about that.
Try another folder to copy. Or figure out were it is with copying and try to copy one file manually to see what happens