December 9, 20232 yr Hello, new user here ! I recently installed Unraid and was working great. My layout is a single 14 tb HDD (zfs) in the array and 2*2tb ssd in a zfs mirror pool. No parity drive for the moment. The 14tb spinner in the array was working fine for approx. a month. I have another 2 14tb HDD as unassigned for the time being, just lying there, recently finished pre-clearing them. What happened: I stopped the array and upon restarting it I get the error "Unmountable: Unsupported partition layout" in the single 14tb HDD in the array. Disk looks fine in SMART and was working fine also until the array stop/start. The only thing changed was an array stop-start. I am attaching diagnostics and would appreciate any help ! myserver-diagnostics-20231209-1430.zip
December 9, 20232 yr Author this is the output: root@myserver:~# fdisk -l /dev/sdb Disk /dev/sdb: 12.73 TiB, 14000519643136 bytes, 3418095616 sectors Disk model: ST14000NM001G-2K Units: sectors of 1 * 4096 = 4096 bytes Sector size (logical/physical): 4096 bytes / 4096 bytes I/O size (minimum/optimal): 4096 bytes / 4096 bytes for comparison, the other two (unassigned, just pre cleared) disks: root@myserver:~# fdisk -l /dev/sdc Disk /dev/sdc: 12.73 TiB, 14000519643136 bytes, 3418095616 sectors Disk model: ST14000NM001G-2K Units: sectors of 1 * 4096 = 4096 bytes Sector size (logical/physical): 4096 bytes / 4096 bytes I/O size (minimum/optimal): 4096 bytes / 4096 bytes Disklabel type: dos Disk identifier: 0x00000000 Device Boot Start End Sectors Size Id Type /dev/sdc1 1 4294967295 4294967295 16T 0 Empty root@myserver:~# fdisk -l /dev/sdd Disk /dev/sdd: 12.73 TiB, 14000519643136 bytes, 3418095616 sectors Disk model: ST14000NM001G-2K Units: sectors of 1 * 4096 = 4096 bytes Sector size (logical/physical): 4096 bytes / 4096 bytes I/O size (minimum/optimal): 4096 bytes / 4096 bytes Disklabel type: dos Disk identifier: 0x00000000 Device Boot Start End Sectors Size Id Type /dev/sdd1 1 4294967295 4294967295 16T 0 Empty Edited December 9, 20232 yr by kokni
December 9, 20232 yr Community Expert Disk1 has no partition, something likely clobbered the MBR, before rebooting it would still use in-memory structures and everything continues to work normally, any idea what could have done that? Did you run any wipefs or similar command?
December 9, 20232 yr Author I have some understanding of the command line, and I can certainly say that I have not run any disk-related commands. The only close hit that I can relate with, is that some days ago I installed the unassigned drives plugin, and using destructive mode precleared the two additional disks. But these operations were performed with the array online and the disk in question (single member of the array) was not an option anyway for me to mess with by mistake. Anyway, that was a zfs formatted drive with unraid. Is there any chance to recover data (it was nearly full) or is it a write off? If it matters, the disk in question was previously on a QNAP QTS 5.0 system. I removed it, installed it in a vanilla Unraid system, assigned it to the array, formatted through the GUI and used it. At that time I did not know about plugins, pre clearing etc. Could it be that the partitions from QNAP QTS were carried over and hence the problem now or am I talking nonsense?
December 9, 20232 yr Community Expert You may be able to recover the old partition with testdisk or by recreating it, testdisk would be safer, recreating the Unraid default partition layout *should* also work, but cannot guarantee, you can clone the disk with dd first to play it safer, then: sgdisk -o -a 8 -n 1:32K:0 /dev/sdX Then re-start the array, I think a reboot won't be needed but try if it still doesn't work, if it still complains about the layout the disk should now at least be mountable with UD.
December 9, 20232 yr Community Expert 29 minutes ago, kokni said: The only close hit that I can relate with, is that some days ago I installed the unassigned drives plugin, and using destructive mode precleared the two additional disks. Unlikely that would have done it, but assuming no diags before the reboot no way to be certain.
December 11, 20232 yr Author Hello Jorge, many thanks for your help. while waiting for dd to finish a clone I did some testing: 1) I recreated with a spare disk the exact steps, the resulting EFI GPT partition table is: Disk /dev/sdd - 14 TB / 12 TiB - CHS 212766 255 63 Current partition structure: Partition Start End Size in sectors 1 P Linux filesys. data 8 3418095610 3418095603 So, this is what I should try to recreate I guess. 2) On the *failed* disk, testdisk shows Disk /dev/sdb - 14 TB / 12 TiB - CHS 212766 255 63 Current partition structure: Partition Start End Size in sectors Bad GPT partition, invalid signature. Trying alternate GPT 1 P Linux filesys. data 8 3418095610 3418095603 but the next step takes ages and more alarmingly, it fins partitions from the previous raid5 membership on a qnap system. I also run gdisk. It finds the following: GPT fdisk (gdisk) version 1.0.9.1 Caution: invalid main GPT header, but valid backup; regenerating main header from backup! Warning: Invalid CRC on main header data; loaded backup partition table. Warning! Main and backup partition tables differ! Use the 'c' and 'e' options on the recovery & transformation menu to examine the two tables. Warning! Main partition table CRC mismatch! Loaded backup partition table instead of main partition table! Warning! One or more CRCs don't match. You should repair the disk! Main header: ERROR Backup header: OK Main partition table: ERROR Backup partition table: OK Partition table scan: MBR: not present BSD: not present APM: not present GPT: damaged Found invalid MBR and corrupt GPT. What do you want to do? (Using the GPT MAY permit recovery of GPT data.) 1 - Use current GPT 2 - Create blank GPT Selecting "Use current GPT" I get Partition GUID code: 0FC63DAF-8483-4772-8E79-3D69D8477DE4 (Linux filesystem) Partition unique GUID: 89773B4D-0457-4CD3-AE4D-A6CE3B7B16D1 First sector: 8 (at 32.0 KiB) Last sector: 3418095610 (at 12.7 TiB) Partition size: 3418095603 sectors (12.7 TiB) Attribute flags: 0000000000000000 Partition name: '' Which looks the same as testdisk. I do not understand well the testdisk program, and the old (raid5) partitions that appear scare me. So, the question: use the command suggested by you above, or use gdisk and try to copy the backup to the main (described here) ?
December 11, 20232 yr Community Expert 33 minutes ago, kokni said: or use gdisk and try to copy the backup to the main I would try this first, then if still the same recreate the partition with that command.
December 12, 20232 yr Author I worked with gdisk, I restored the backup to the main entry and everything is OK with no data loss. Thanks for your guidance !!! However, at this point I have zero faith in how Unraid handled the installed disks, so as a precaution I checked my mirrored zfs pool, which seemingly works for the moment. But see below: fdisk -l /dev/nvme0n1 Disk /dev/nvme0n1: 1.86 TiB, 2048408248320 bytes, 4000797360 sectors Disk model: TEAM TM8FPD002T Units: sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disklabel type: dos Disk identifier: 0x00000000 Device Boot Start End Sectors Size Id Type /dev/nvme0n1p1 2048 4000797359 4000795312 1.9T 83 Linux and root@myserver:~# gdisk /dev/nvme0n1 GPT fdisk (gdisk) version 1.0.9.1 Partition table scan: MBR: MBR only BSD: not present APM: not present GPT: not present *************************************************************** Found invalid GPT and valid MBR; converting MBR to GPT format in memory. THIS OPERATION IS POTENTIALLY DESTRUCTIVE! Exit by typing 'q' if you don't want to convert your MBR partitions to GPT format! *************************************************************** Warning! Secondary partition table overlaps the last partition by 33 blocks! You will need to delete this partition or resize it in another utility. Disk /dev/nvme0n1 - 2048 GB / 1907 GiB - CHS 1953514 64 32 Current partition structure: Partition Start End Size in sectors No ext2, JFS, Reiser, cramfs or XFS marker 1 P Linux 1 0 1 1953514 21 16 4000795312 1 P Linux 1 0 1 1953514 21 16 4000795312 Warning: Bad starting sector (CHS and LBA don't match) No partition is bootable @JorgeB this was a pair of nvme's, installed and assigned as a zfs mirrored pool, formatted through the GUI and that's it. How on earth did the above happened, which is a disaster waiting to happen? I really do not understand what destroyed all the GPT/MBR entries.
December 12, 20232 yr Community Expert The fdisk output for the pool looks normal, were these brand new devices or have they been used before elsewhere?
December 12, 20232 yr Author 45 minutes ago, JorgeB said: The fdisk output for the pool looks normal, were these brand new devices or have they been used before elsewhere? It looks deceptively normal, but please note that it is disktype=dos instead of disktype=gpt and moreover testdisk recognizes "intel" partition table type instead of "efi-gpt". Indeed, all drives were previously on a QNAP QTS 5.0 system, which I believe is the root of all problems. I had the (wrong?) immersion that bringing the drives into Unraid, assigning them to arrays/pools and formatting them through the GUI would also take care of the correct MBR/GPT. Apparently, this sequence causes problems with drives coming from a QNAP QTS5.0 system? edit: also, not sure about the 512 sector being optimal, investigating this now. Edited December 12, 20232 yr by kokni
December 12, 20232 yr Community Expert 27 minutes ago, kokni said: It looks deceptively normal, but please note that it is disktype=dos instead of disktype=gpt and moreover testdisk recognizes "intel" partition table type instead of "efi-gpt". That's normal, Unraid only creates GPT partitions for devices >2TB.
December 13, 20232 yr Author Ah, thanks for this clarification. How should I proceed then, leave it as is despite the warnings or destroy the pool and start over? Also, I hope to upgrade to 4gb mirror soon. When I will replace the first 2tb with a 4tb, will the zfs pool resilver without problems with one disk being MBR and the new GPT?
December 14, 20232 yr Community Expert 12 hours ago, kokni said: How should I proceed then, leave it as is despite the warnings or destroy the pool and start over? I'd leave it, warnings might be because gdisk is to be used with GPT partitions. 12 hours ago, kokni said: When I will replace the first 2tb with a 4tb, will the zfs pool resilver without problems with one disk being MBR and the new GPT? Correct.
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.