CF Failure?

[Aussie101]

Hi.

My Alix3D is sick...

We had a power outage. After the power was restored the Alix came up with errors on /dev/hda3 which is /data

I connected a keyboard and ran fsck manually. Lost the data, but found some files (graphs etc) in /data/Lost+Found/

Recreated the /data disrectories and meteo happily started logging and updating wunderground again.

This morning:

Code: Select all

Buffer I/O error on device hda3, logical block 3958107115
lost page write due to I/O error on hda3
Aborting journal on device hda3.
ext3_abort called.
EXT3-fs error (device hda3): ext3_journal_start_sb: Detected aborted journal
Remounting filesystem read-only
attempt to access beyond end of device
hda1: rw=1, want=18190449728, limit=1574307
Buffer I/O error on device hda1, logical block 2273806215
lost page write due to I/O error on hda1
Aborting journal on device hda1.
ext3_abort called.
EXT3-fs error (device hda1): ext3_journal_start_sb: Detected aborted journal
Remounting filesystem read-only
alarm[9142]: segfault at 00000046 eip b7ebc66d esp bfa06b70 error 4

Doesn't look good. I can log in, and the wunderground logging is running ok, but the meteohub web interface says:

Code: Select all

Error file open /home/meteohub/meteohub.conf

I think if I reboot it now, it will not come back up, repairs to the hda1 will trash the linux system.

I've only get a spare slow Kingston camera CF - probably MLC, have ordered a new SLC card, but I somehow doubt that it will arrive in time.

Do you think this CF is broken or will a blank and re-install on this CF solve my problems?

Cheers,

Michael

[Aussie101]

Talking to myself here...

Anyway, system is still up and logging to wunderground. Even though /data is not writable, it has not stopped the core station reading and logging to wunderground. This is quite remarkable

I thought to try a jumpdrive to copy the /home/meteohub but sadly, cannot get the usb_storage module to load - it's not there at all, nor the associated scsi module. Too bad, I might have been able to salvage something... Note for Admin, this would be worth including in future builds. Also might be worth having a mountpoint for usbstorage in tmpfs in case anyone else ever has this issue.

smb is down, haven't investigated the webserver yet (still running) to see if I can place something in it's path, but if it is serving up from hda1 or hda3 then that won't work either. rsync to try as well, but not hopeful on that.

Tried nc, no joy. it errors immediately, perhaps because it creates a temp file somewhere other than tmpfs...

If anyone has other ideas, be happy to hear...

Michael

[Aussie101]

Woot!

Code: Select all

# rsync -uav rsync://root@192.168.1.245/config myconfig

receiving file list ... done
created directory myconfig
./
.bash_logout
.bash_profile
.bashrc
.cleanup
a850
activation
alarm
alive
checketh0
clientraw.conf
[..] (continues)..

sent 1876 bytes  received 1932207 bytes  3868166.00 bytes/sec
total size is 1926100  speedup is 1.00

That was unexpected

Also managed to get the graph definitions from /data by the same method. Historical data is lost, but I do have the long term store at wunderground including pre-metohub data.

Now, I'll just wait for the Alix to crash. Hopefully the new CF card will arrive in time.

[Aussie101]

Had another power outage a day or so ago.

Alix booted up clean, no errors. After a few hours, it logged an error for the journal file 'attempt to access beyond the end of device' or something like that and then remounted hda3 as readonly. New CF cards turned up today, so I will re-install at next available opportunity.

If I re-image the latest meteohub to a new CF, how do I transfer my settings over? Is it simply a case of rsyc the ~/config and /data directories and reboot?

Will I need to re-activate the license?

[admin]

yes, new activation will be necessary. when you have the credentials of the old system (system id and activation code) this can be done easily online.

most settings are in meteohub.conf, schedules for graph uploads are in crontab.

[Aussie101]

And its installed and online again.

I restored meteohub.conf and any modified files in /data and rebooted, found my original system info and relicensed, and it all started working normally again.

Cheers,

Michael