I've been running version: 5.0-rc5 for a few months now without a hitch. Problems began a few days ago when the server crashed. After hard-restart and parity rebuilding I noticed that a drive was 99% full.

 

I received one of these for each share:

 

Tower afpd[7193]: volume "Music" does not support Extended Attributes, using ea:ad instead

Tower shfs/user: shfs_setxattr: setxattr: user.org.netatalk.supports-eas.pwpuO1 /mnt/disk1/Pictures (28) No space left on device

 

I had set it to high-water split level 2, but did not set minimum free space. My bad. I freed up space on disk1 by moving chunks internally to disk3 and those errors went away. However, when the mover ran overnight the system crashed again. I could not obtain a log but I took a photo of the screen:

 

 

After doing some research in the forums, I thought this might be a mover issue. I restarted and ran another parity check. All good. There was still alot of data left on the cache drive so I manually deleted those files.  I was able to run the mover manually and the server did not crash. I ran a memtest as well just to be sure it was not that.

 

 

All seemed well until today when new problems arose. These cdid2.db database errors knocked the array off line.

Nov 30 01:05:27 Tower emhttp: shcmd (438): cp /etc/avahi/services/afp.service- /etc/avahi/services/afp.service

Nov 30 01:05:27 Tower avahi-daemon[3201]: Files changed, reloading.

Nov 30 01:05:27 Tower avahi-daemon[3201]: Service group file /services/afp.service changed, reloading.

Nov 30 01:05:27 Tower emhttp: shcmd (439): cp /etc/avahi/services/smb.service- /etc/avahi/services/smb.service

Nov 30 01:05:27 Tower avahi-daemon[3201]: Files changed, reloading.

Nov 30 01:05:27 Tower avahi-daemon[3201]: Service group file /services/smb.service changed, reloading.

Nov 30 01:05:27 Tower emhttp: shcmd (440): /usr/local/sbin/emhttp_event svcs_restarted

Nov 30 01:05:27 Tower emhttp_event: svcs_restarted

Nov 30 01:05:28 Tower avahi-daemon[3201]: Service "Tower" (/services/afp.service) successfully established.

Nov 30 01:05:28 Tower avahi-daemon[3201]: Service "Tower-SMB" (/services/smb.service) successfully established.

Nov 30 01:15:16 Tower shfs/user: shfs_create: create_path: /mnt/cache/Pictures Pictures/.AppleDB/__db.001 /mnt/disk1/Pictures ( Exec format error

Nov 30 01:15:16 Tower cnid_dbd[5538]: error opening DB environment after recovery: Exec format error

Nov 30 01:15:16 Tower afpd[5140]: read: Connection reset by peer

Nov 30 01:15:17 Tower cnid_dbd[5541]: error retrieving value from cnid2.db: Invalid argument

Nov 30 01:15:17 Tower cnid_dbd[5541]: dbif_getversion: database error

Nov 30 01:15:17 Tower afpd[5140]: read: Connection reset by peer

Nov 30 01:15:18 Tower cnid_dbd[5546]: error retrieving value from cnid2.db: Invalid argument

Nov 30 01:15:18 Tower cnid_dbd[5546]: dbif_getversion: database error

Nov 30 01:15:18 Tower afpd[5140]: read: Connection reset by peer

Nov 30 01:15:19 Tower cnid_metad[5552]: Multiple attempts to start CNID db daemon for "/mnt/user/Pictures" failed, wiping the slate clean...

Nov 30 01:15:19 Tower cnid_dbd[5552]: main: too many CNID db opening attempts, wiping the slate clean

 

Did I, in my cleanup efforts, manage to delete data off some of the disks? So I started to panic and, not thinking clearly, I very stupidly disabled my SMB security shares thinking I was just getting rid of the SMB sharing protocol (which I at one point I thought was a cause). And of course now the shares are gone.  Parity is still good and I can access my files on the individual disks.  And now the cache drive is coming up unformatted.

 

So, how do I get my shares back and how do I fix these database errors.. I really messed something up here and I don't dare make any further guesses as to what is wrong. I've guessed enough. I hope you can help.

 

 

syslog-20121130.zip

The errors are caused by Netatalk. Because of the way AFP works, Netatalk ceates some database files within the filesystem. You will have to or disable AFP or set a min free space to each of shares, 20MB should be enough.

Thanks for the response.

 

I neglected to mention that I set min free space after i realized the problem with filled disk. These errors came up after I set min free space on all of the shares. It's all MAC's around here, won't I take a performance hit with SMB?

 

I will try what you suggest.

 

You could try this: http://lime-technology.com/forum/index.php?topic=3999.0

And see BIOS settings in my sig.

Thank you for the reply. I am amazed at how responsive this forum ease. Major Kudos to all of  you.

 

The alloc errors went away after the mover problem stopped. I will check this and the BIOS as well when I get home tonight.

 

Will the shares come back if I reset the

diskExport.1="e"

in share.cfg? I am thinking that's what caused the share to disappear.