February 23, 201115 yr I'm fairly new to unRAID having only setup my system within the last month. The last couple of days I started copying over several TB of data to start filling the system. Currently I have a 1 parity and 2 data drive setup. Everything was going wonderful until my writes to a user share started failing. After some testing on the command line I found that I could write a file to /mnt/disk1 fine but /mnt/disk2 would take up to a minute to write a 10 character text file. At first I thought disk2 was bad and scanned the logs for errors but found none. I even checked the SMART status with no indication of anything bad. After suspecting that maybe the parity drive had an issue I replaced it with a standby drive and now I seem to be able to write once again to the array. The parity rebuild is still going though, I don't know if that will change things once it has completed. I suppose then that my question would be is it possible that my parity drive was in fact bad, despite the lack of errors? I've searched for quite a while and can't seem to find anyone that has had a similar issue, which scares me and makes me wonder if the problem is still here and will reappear later.
February 23, 201115 yr please attach a zipped syslog so we can check things ... without syslog we are speculating as much as you are...
February 23, 201115 yr Author please attach a zipped syslog so we can check things ... without syslog we are speculating as much as you are... Here's my current syslog. This is once the parity drive has been replaced and is currently rebuilding. syslog.zip
February 23, 201115 yr Can you write to disk2 currently? It does not matter if parity is still rebuilding.
February 23, 201115 yr Author Can you write to disk2 currently? It does not matter if parity is still rebuilding. Yes, after swapping out the parity drive I am able to write to disk2. I realized in my original post I forgot to mention that during this I also attempted to swap out disk2 itself and continued to have the same issue. During each step I also checked out the syslog and never saw any indication of errors.
February 23, 201115 yr Author Well now I really don't know what to try. Once the parity finished rebuilding the problem immediately reappeared. Once again I can write to disk1 but not disk2 at all. It just hangs even trying to create a few character text file with vi. Still nothing in syslog.
February 23, 201115 yr Post a syslog taken after the problem happens. We need the whole thing zip it if needed.
February 23, 201115 yr Author Post a syslog taken after the problem happens. We need the whole thing zip it if needed. Current syslog attached. syslog.zip
February 23, 201115 yr While it does show a problem, I'm finding the log inconclusive. Since you already changed the swap drive, I would replace disk2 with your prior parity drive and let it rebuild.
February 23, 201115 yr Author I haven't done the disk replacement yet since I'm waiting for a preclear to run but as if to spite me I can no longer do an ls on a specific folder within /mnt/disk1, it just hangs. Yet I am able to list other folders on disk1 without trouble. This may be far out in left field but is it possible that disk1 is actually bad and when I try to write to disk2 the hang is due to not being able to read disk1 for parity or does it not work that way? This might also explain why it was working when the parity drive was being rebuilt. Then again if that was the case it shouldn't be able to recreate the parity drive at all. All of this almost makes me want to go back to my Drobo despite it's faults. :'(
February 24, 201115 yr Author You did check to make sure your cables are connected and not loose right? Yep, one of the first things I did was to re-seat every cable and drive when the problem first appeared.
February 24, 201115 yr I haven't done the disk replacement yet since I'm waiting for a preclear to run but as if to spite me I can no longer do an ls on a specific folder within /mnt/disk1, it just hangs. Yet I am able to list other folders on disk1 without trouble. This may be far out in left field but is it possible that disk1 is actually bad and when I try to write to disk2 the hang is due to not being able to read disk1 for parity or does it not work that way? This might also explain why it was working when the parity drive was being rebuilt. Then again if that was the case it shouldn't be able to recreate the parity drive at all. All of this almost makes me want to go back to my Drobo despite it's faults. :'( No. Disk1 is not involved in reads or writes to disk2.
February 24, 201115 yr Author I haven't done the disk replacement yet since I'm waiting for a preclear to run but as if to spite me I can no longer do an ls on a specific folder within /mnt/disk1, it just hangs. Yet I am able to list other folders on disk1 without trouble. This may be far out in left field but is it possible that disk1 is actually bad and when I try to write to disk2 the hang is due to not being able to read disk1 for parity or does it not work that way? This might also explain why it was working when the parity drive was being rebuilt. Then again if that was the case it shouldn't be able to recreate the parity drive at all. All of this almost makes me want to go back to my Drobo despite it's faults. :'( No. Disk1 is not involved in reads or writes to disk2. I'm just cursed then. I'll post how it goes once I replace disk2.
February 24, 201115 yr Author Please post your system specs. COOLER MASTER Centurion 590 ASUS M4A785-M AM3/AM2+/AM2 AMD 785G HDMI Micro ATX AMD Motherboard CORSAIR Builder Series CX430 CMPSU-430CX 430W ATX12V Active PFC Power Supply AMD Sempron 140 Sargas 2.7GHz 1MB L2 Cache Socket AM3 45W Single-Core Processor SDX140HBGQBOX Kingston 2GB 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Desktop Memory Model KVR800D2N6/2G SATA2 Serial ATA II PCI-Express RAID Controller Card (Silicon Image SIL3132) 2 Port SATA Serial ATA PCI RAID Controller Card - Silicon Image 9x Mobile Rack for 3.5 inches SATA HDD - Open Frame 3x Western Digital 2 TB Caviar Green SATA Intellipower 64 MB Cache Bulk/OEM Desktop Hard Drive WD20EARS Pretty much the 9 drive system from Raj's prototype designs thread.
February 24, 201115 yr You could try reiserfsck --check /dev/mdx where x=1 then x=2. You can do this during rebuild.
February 24, 201115 yr I was having the same issue, with writing to one drive was really slow and now it's working fine. I made sure my cables were tight and I did a memory test and now everything is working fine.
February 24, 201115 yr You could try reiserfsck --check /dev/mdx where x=1 then x=2. You can do this during rebuild. after stopping samba and un-mounting the disk being checked
February 24, 201115 yr You could try reiserfsck --check /dev/mdx where x=1 then x=2. You can do this during rebuild. after stopping samba and un-mounting the disk being checked What does this do?
February 24, 201115 yr Are you using user shares when you write to your array? If so, what are your share settings (minimum space, split level, allocation method)? How full is the "failing" disk?
February 24, 201115 yr Author Are you using user shares when you write to your array? If so, what are your share settings (minimum space, split level, allocation method)? How full is the "failing" disk? I've placed the original parity drive back into the array as disk2 and it's now rebuilding. I'm also running the reiserfsck check mentioned by dgaschk so I'll update with what happens there as well. I am using a user share with 16GB min space, split level 2 and high-water. The problem disk is a roughly half full 2TB drive. Although after I noticed the initial problem on my user share I began debugging using the console and writing directly to the /mnt/diskX partitions in order to eliminate the share as a problem.
February 24, 201115 yr You could try reiserfsck --check /dev/mdx where x=1 then x=2. You can do this during rebuild. after stopping samba and un-mounting the disk being checked I realized that this may not be possible. Can you unmount during a rebuild?
Archived
This topic is now archived and is closed to further replies.