So since May I had configured FreeNas to have 2 pools. One pool consisting of 10 x WD Red 3TB drives in a Raidz3 configuration and a second pool of 2 x 256GB Samsung 850 Pro drivers in a mirror. Everything was working fine until last week. Last week I got the "The volume Pool (ZFS) state is ONLINE: One or more devices has experienced an error resulting in data corruption. Applications may be affected." error which I then ran a scrub on and saw there were a lot of permanent errors on my files which is mostly media files.
The first thing I did was run SMART tests with the "smartctl -x /dev/daX" command and didn't really see any jump out to me. Strange thing was that the error came up for both Pools of disks. After searching thru the forums I saw people thinking it might be bad ram so I ran a memtest on my ram and let it complete a full pass which came back with no errors. Next thing I did was I destroyed my pool and recreated it as a striped mirrors and proceeded to move just one media file on the newly created pool. I ran a scrub after it and it had listed that one file as a permanent error. After deleting the file and rerunning the scrub it would come back with no permanent errors. I then thought it was my media that was sitting on my old htpc which I then did a chkdsk on it came back without any errors. Next I thought it was my network somehow corrupting the files when I was transferring the media from my old htpc to my newly created pool. I did a checksum on a file on both the htpc and after being copied to the pool and they matched.
I then realized it might be something to do with the fact my onboard LSI controller had firmware 19 compared to the driver 20 of freenas. This error came up when I first installed FreeNas and I thought it wasn't a big deal until I started reading about how it might mess stuff up. So I flashed it to firmware to version 20. Thinking that had to had been the problem I continued to transfer 1.8TB of movies onto the newly created stripped mirror pool. As soon as the transfer was complete I ran a scrub and it came back with no errors. I continued to transfer all my content back on the pool and ran another scrub just to be safe and again no errors.
I thought I had fixed it. This morning I clicked on the green alert button and noticed I had received the "CRITICAL: The volume Pool (ZFS) state is ONLINE: One or more devices has experienced an error resulting in data corruption. Applications may be affected." without freenas emailing me which it did the previous times also strange that it was already unchecked. So I proceeded to run a scrub this morning and it found some errors. I'm running out of ideas as to why its happening im thinking its not my drives health that's the problem as they were just purchased in May.
My setup is a Supermicro X8DT6-F motherboard with the onboard LSI controlled flashed to IT mode and passthrough to my FreeNas VM with 4 cpus and 32GB of ram. My host is running Ubuntu Server 15.10 and im using KVM to manage the VM's. My ram is ECC and the cpus my host has is 2 x Xenon E5650 six core processors.
The first thing I did was run SMART tests with the "smartctl -x /dev/daX" command and didn't really see any jump out to me. Strange thing was that the error came up for both Pools of disks. After searching thru the forums I saw people thinking it might be bad ram so I ran a memtest on my ram and let it complete a full pass which came back with no errors. Next thing I did was I destroyed my pool and recreated it as a striped mirrors and proceeded to move just one media file on the newly created pool. I ran a scrub after it and it had listed that one file as a permanent error. After deleting the file and rerunning the scrub it would come back with no permanent errors. I then thought it was my media that was sitting on my old htpc which I then did a chkdsk on it came back without any errors. Next I thought it was my network somehow corrupting the files when I was transferring the media from my old htpc to my newly created pool. I did a checksum on a file on both the htpc and after being copied to the pool and they matched.
I then realized it might be something to do with the fact my onboard LSI controller had firmware 19 compared to the driver 20 of freenas. This error came up when I first installed FreeNas and I thought it wasn't a big deal until I started reading about how it might mess stuff up. So I flashed it to firmware to version 20. Thinking that had to had been the problem I continued to transfer 1.8TB of movies onto the newly created stripped mirror pool. As soon as the transfer was complete I ran a scrub and it came back with no errors. I continued to transfer all my content back on the pool and ran another scrub just to be safe and again no errors.
I thought I had fixed it. This morning I clicked on the green alert button and noticed I had received the "CRITICAL: The volume Pool (ZFS) state is ONLINE: One or more devices has experienced an error resulting in data corruption. Applications may be affected." without freenas emailing me which it did the previous times also strange that it was already unchecked. So I proceeded to run a scrub this morning and it found some errors. I'm running out of ideas as to why its happening im thinking its not my drives health that's the problem as they were just purchased in May.
My setup is a Supermicro X8DT6-F motherboard with the onboard LSI controlled flashed to IT mode and passthrough to my FreeNas VM with 4 cpus and 32GB of ram. My host is running Ubuntu Server 15.10 and im using KVM to manage the VM's. My ram is ECC and the cpus my host has is 2 x Xenon E5650 six core processors.
