My unraid has been working flawlessly for months with 8 drives + parity. The other day it suddenly stopped accepting connections, although the web interface was still working. I installed unmenu to see if it could provide any insights for me, and it reported (in big, angry looking red letters), that "SAMBA is STOPPED, Shared drives will not be visible on the LAN." I did a complete clean re-installation of unraid which seemd to work fine, although it reported 28,000 sync errors! (it took a few days to munch its way through them). Anyway I got home today and the array seemed to be working again until I tried a reboot, and the same problem has occured again. I have attached the syslog and would be really grateful if anyone can shed some light on this for me. 

My unraid has been working flawlessly for months with 8 drives + parity. The other day it suddenly stopped accepting connections, although the web interface was still working. I installed unmenu to see if it could provide any insights for me,

It does provide some diagnostic tools that would help.  You could easily examine the syslog, or export it for analysis here... or look at a process list to see what it running

and it reported (in big, angry looking red letters), that "SAMBA is STOPPED, Shared drives will not be visible on the LAN."

That indicated that the smbd process was not running.

  I did a complete clean re-installation of unraid which seemd to work fine,

Did you think this is like a windows OS.  The equivalent to what you said in automotive terms would be, the ash-tray needed emptying, so I rebuilt the engine and painted the body in a new color. (but dang, it, the ash-tray was still full)

although it reported 28,000 sync errors! (it took a few days to munch its way through them).

There is an underlying issue that caused those errors...  I do not have a feeling they are corrected... not until you perform a second full parity check and it comes up with zero errors.  But, don't worry about that just yet, you havr other issues now as evidenced by your attached syslog.

Anyway I got home today and the array seemed to be working again until I tried a reboot, and the same problem has occured again.

What happened when you tried to reboot?  You did not say.

I have attached the syslog and would be really grateful if anyone can shed some light on this for me. 

This is the syslog immediately after rebooting as far as I can tell.

 

I can see a whole series of error messages regarding your flash drive.  The file-system on it is apparently corrupt:

un 11 12:55:58 Tower kernel: FAT: Filesystem panic (dev sdd1)

Jun 11 12:55:58 Tower kernel:    fat_get_cluster: invalid cluster chain (i_pos 131215)

Jun 11 12:55:58 Tower kernel:    File system has been set read-only

Jun 11 12:55:58 Tower kernel: FAT: Filesystem panic (dev sdd1)

Jun 11 12:55:58 Tower kernel:    fat_get_cluster: invalid cluster chain (i_pos 131215)

Jun 11 12:55:58 Tower kernel: FAT: Filesystem panic (dev sdd1)

Jun 11 12:55:58 Tower kernel:    fat_get_cluster: invalid cluster chain (i_pos 131215)

Jun 11 12:55:58 Tower kernel: FAT: Filesystem panic (dev sdd1)

Jun 11 12:55:58 Tower kernel:    fat_get_cluster: invalid cluster chain (i_pos 131215)

Jun 11 12:55:58 Tower kernel: FAT: Filesystem panic (dev sdd1)

Jun 11 12:55:58 Tower kernel:    fat_get_cluster: invalid cluster chain (i_pos 131215)

Jun 11 12:55:58 Tower kernel: FAT: Filesystem panic (dev sdd1)

Jun 11 12:55:58 Tower kernel:    fat_get_cluster: invalid cluster chain (i_pos 131215)

Jun 11 12:55:58 Tower kernel: FAT: Filesystem panic (dev sdd1)

Jun 11 12:55:58 Tower kernel:    fat_get_cluster: invalid cluster chain (i_pos 131215)

 

You probably need to put it in a windows box and run scandisk on it.    Note, it has been set as read-only... That means when you reboot you will not be able to save the fact that a clean shutdown occurred, and you will be forced to do a full parity check upon restart of the array.  Once it has a valid file-system, you can reload unRAID on it.  (something you will almost NEVER need to do unless it gets corrupted by something... I've have an unRAID box here for years, and I've never had to reformat the flash drives...  I've upgraded the OS tons of times, but that generally involves replacing two files on the flash drive and rebooting... not reformatting and re-loading.  Perhaps when you re-loaded it, you did not safely eject it... and that is where it got the corruption.  Who knows?)

 

Usually, if smbd (SAMBA) has been killed, it is because the system log filled with error messages and to make room in RAM Linux starts terminating processes it does not think are in use at the moment.  smbd probably got killed...  If you had gotten a copy of the syslog BEFORE you rebooted, we would have had more clues.  Now, you'll have to either monitor the syslog for errors, (after fixing the flash drive FAT filesystem corruption) or wait till it kills smbd once more.

 

If by chance the flash drive is not corrupted, then you have probably have a memory problem. (Either a bad strip, or the voltage, timing, or clock speed set incorrectly)  That also can easily cause the thousands of parity errors you saw.  They are not normal, and the cause is still unknown.  A bad flash drive file-system would not be their cause.

 

Joe L.

Thank you very much for the prompt and comprehensive response Joe. I did as you suggested and ran the USB key through Checkdisk. It did indeed find a corrupt file called 'secrets.tdb', which it fixed. On re-starting unraid it now seems to be working again fine. I've started another parity check which will hopefully report no errors.

 

Thanks again for your help. 

 

Robert

Fanfckntasic, I just used this thread to a problem I was having where I was deleting files from the server and was taking awfully long, stopped the array and it got stuck on the unmounting. I then 'panicked' and cold-shut down the server and since starting it up everything looked fine but couldn't then connect to my shares. I tried to then telnet to it and do a reiserfsck and it error'ed out on the secrets file, searched here and found this thread. Did what Joe suggested and put the flash drive back into the server, all good again. THANK YOU THANK YOU THANK YOU!