September 3, 20178 yr Upgraded to 6.4.0-rc8q tonight. After rebooting, my cache drive will not mount- status is Unmountable: Invalid partition If I remove it from the cache and mount it with Unassigned Devices, it works just fine. It's formatted btrfs and is a single SSD. Any advice or further data that would be helpful?
September 3, 20178 yr Upgraded to 6.4.0-rc8q tonight. After rebooting, my cache drive will not mount- status is Unmountable: Invalid partition If I remove it from the cache and mount it with Unassigned Devices, it works just fine. It's formatted btrfs and is a single SSD. Any advice or further data that would be helpful?I realize that you only have one assigned, but how many cache slots are available when the array is stopped?Sent from my SM-G930P using Tapatalk
September 3, 20178 yr Author Just the 1 slot is assigned, and 1 selected from drop down. No empty slots.
September 3, 20178 yr Author And in case it helps, I misspoke about the drive- was thinking of the wrong server. Don't know if it will make a difference, but it's actually a 3 drive RAID-5 managed by an Areca 1883ix and presented to unRAID as a single 1tb drive.
September 3, 20178 yr 13 minutes ago, eoh7678 said: Upgraded to 6.4.0-rc8q tonight. After rebooting, my cache drive will not mount- status is Unmountable: Invalid partition If I remove it from the cache and mount it with Unassigned Devices, it works just fine. It's formatted btrfs and is a single SSD. Any advice or further data that would be helpful? diagnostics, always post diagnostics......
September 3, 20178 yr Author Diagnostics attached. Sorry- haven't posted here much as of yet. Only bought Pro a few weeks ago. Appreciate the fast responses, though! home-srv-ur-1-diagnostics-20170902-1956.zip
September 3, 20178 yr Author In the meantime, would creating a symlink from the mountpoint of the cache using unassigned devices to /mnt/cache maintain the chain and get things going as a workaround?
September 3, 20178 yr I have the same problem. I replaced my cache disk on rc7 and it would not format. I had to format it with the UD plugin and it mounted with rc7, but won't mount with rc8. I'm now moving the data off the cache under rc7 and the I'll format the cache under rc8 and copy data back. I think the starting sector was 2048 when partitioned with UD and unRAID does not accept that under rc8.
September 3, 20178 yr Author Ugh, that's what I was afraid of. I've already started pulling the data from the drive figuring that'd be my solution of last resort.
September 3, 20178 yr 9 minutes ago, eoh7678 said: Ugh, that's what I was afraid of. I've already started pulling the data from the drive figuring that'd be my solution of last resort. It's just a matter of time. You realize you can set the shares that are 'Cache Only' to 'Yes', stop Docker and VMs and start the mover and the data will be moved off the cache. Format the cache disk and then set the shares to cache 'Prefer' and start the mover again and the data will be moved back. The set the shares to your preferred cache method. EDIT: I have about 400GB to move. Ugh, it's going to take a while. Edited September 3, 20178 yr by dlandon
September 3, 20178 yr Author The cache drive won't mount at all, so I can't do that. I can only mount it using Unassigned Devices. Unless the symlink idea mentioned earlier would trick ur long enough for me to do it that way. And I hear you- I've got about 500GB. Edited September 3, 20178 yr by eoh7678
September 3, 20178 yr 3 minutes ago, eoh7678 said: The cache drive won't mount at all, so I can't do that. I can only mount it using Unassigned Devices. Unless the symlink idea mentioned earlier would trick ur long enough for me to do it that way. Go back to rc7 and move the data off the cache.
September 3, 20178 yr Author Just now, dlandon said: Go back to rc7 and move the data off the cache. That sounds promising. Will search for process now. Thanks for the tip.
September 3, 20178 yr 1 minute ago, eoh7678 said: That sounds promising. Will search for process now. Thanks for the tip. Copy files from the previous folder on the flash to the flash and reboot. You'll go back to rc7.
September 3, 20178 yr Author Thanks for that! Will try after I download stuff from the cache drive. I know my luck- if I don't backup first, I'll revert, reboot, and the cache drive will somehow be wiped. I've got a nice fast Mellanox 10gbit card, so I get close to 600MB/s downloads from the cache drive. It won't take too long. Edited September 3, 20178 yr by eoh7678
September 3, 20178 yr 6 hours ago, dlandon said: I think the starting sector was 2048 when partitioned with UD and unRAID does not accept that under rc8. RC8 has now built-in support for 4K sector disks, not sure if related but it may interfere with a 2K boundary.
September 3, 20178 yr 4 hours ago, bonienl said: RC8 has now built-in support for 4K sector disks, not sure if related but it may interfere with a 2K boundary. Probably. I had to use UD to format the cache drive under rc7 because I got some sort of an error that prevented it from being formatted. I can't remember the error. UD tends to use standard partition defaults when partitioning disks which are not necessarily compatible with unRAID. I expected after formatting the disk with UD, that unRAID would not recognize the partition and ask to reformat the disk. unRAID didn't complain and just accepted the existing partition. I used the disk as it was and loaded up the cache. I'm hoping that LT will put some time into partitioning disks once UD is folded into unRAID and possibly standardize formatting in UD so disks can be moved around in unRAID and back and forth with other Linux systems. I suspect it's not that easy to change unRAID partitioning because of the need for a consistent partition offset for parity operations.
September 3, 20178 yr I posted in the main thread as well, but I'm having this issue as well. 2 Parity Drives, 2 Array Drives, 1 cache. Happy to do testing, I can deal without my dockers/vms for a bit. redemption-diagnostics-20170903-1129.zip
September 3, 20178 yr 7 minutes ago, smdion said: I posted in the main thread as well, but I'm having this issue as well. 2 Parity Drives, 2 Array Drives, 1 cache. Happy to do testing, I can deal without my dockers/vms for a bit. redemption-diagnostics-20170903-1129.zip Click on the disk name on the 'Main' page and see if the format is unknown. If it is then your disks are not partitioned to start at sectors 63 or 64. Did you happen to format and partition any disks with UD?
September 3, 20178 yr Author FYI, I just ended up moving the data off the cache in 6.4 rc7, recreated the cache drive under 6.4 rc8 and formatting it XFS, then migrating the data back. Since it's an externally managed RAID-5 of 3 SSDs I didn't see the point in using btrfs.
September 3, 20178 yr 3 hours ago, dlandon said: Click on the disk name on the 'Main' page and see if the format is unknown. If it is then your disks are not partitioned to start at sectors 63 or 64. Did you happen to format and partition any disks with UD? Convo happened in main 6.4 rc8q thread. Edited September 3, 20178 yr by smdion
Archived
This topic is now archived and is closed to further replies.