July 30, 20205 yr Just installed a new WD 8TB RED WD80EFAX hdd in my system. Ran a preclear (zeroing) with the preclear plugin. Stopped the array, added the new disk to the array, started the array, formatted the new disk (as xfs). The overview now shows that 56GB of the 8TB of that new disk is used. Investigating the disk with the terminal shows it contains 1 folder with 16 small files (between 1 and 250KB in size). I know there always is some os overhead, sure. But 56GB? What could be going on here? Edited July 30, 20205 yr by ssh
July 30, 20205 yr Community Expert 25 minutes ago, ssh said: What could be going on here? 25 minutes ago, ssh said: overhead It's normal for newer xfs file systems, they use more space for reflink support and such.
July 30, 20205 yr Author 13 minutes ago, johnnie.black said: It's normal for newer xfs file systems, they use more space for reflink support and such. Okay, thanks Just for my curiosity; does that scale with the total disk size?
July 30, 20205 yr Community Expert 39 minutes ago, ssh said: does that scale with the total disk size? Yep, the larger the disk, the larger the overhead.
September 7, 20205 yr On 7/30/2020 at 1:04 PM, JorgeB said: It's normal for newer xfs file systems, they use more space for reflink support and such. Is it possible to format the old way with reflink=0 on Unraid 6.8.3? I use my NVME cache pool with BTRFS for VM and docker. My array is formatted with XFS for media files mostly and there I do not see much benefits from reflink feature.
September 8, 20205 yr Community Expert 7 hours ago, Forusim said: Is it possible to format the old way with reflink=0 on Unraid 6.8.3? Only if you format the device manually.
September 8, 20205 yr 13 hours ago, JorgeB said: Only if you format the device manually. Are there any special arguments, Unraid is using for formatting before adding to the array? A command line for manual XFS format is much appreciated.
September 9, 20205 yr Community Expert 10 hours ago, Forusim said: Unraid is using for formatting before adding to the array? Format is done after the device is part of the array, so parity can be updated. Default Unraid xfs format is: mkfs.xfs -m crc=1,finobt=1 -f /dev/mdX where X is the disk #, since I don't use xfs for a long time not current on the options, but probably crc=0 would get rid of the extra space, google "mkfs.xfs man page" for the options. Format will likely not work on a mounted filesystem, you'd need to manually unmount, probably easier for you to do this: -stop array -click on the disk you want to format and change it to a different filesystem -start array, that disk won't mount -type the correct mkfs command for the disk on the console (all data on that disk will be deleted) -stop array, change its fs back to xfs/auto -start array, if you used the correct format options it will now use less space.
September 10, 20205 yr @JorgeB Your guide worked out perfectly for me, thanks a lot. I adjusted your command a little bit in regard to sectorsize to match my existing disks: mkfs.xfs -m crc=1,finobt=1,reflink=0 -l su=4096 -s size=4096 -f /dev/mdX Now the overhead dropped from 78 GB to only 11 GB. Edited September 10, 20205 yr by Forusim
June 14, 20224 yr On 9/9/2020 at 4:42 PM, JorgeB said: Format is done after the device is part of the array, so parity can be updated. Default Unraid xfs format is: mkfs.xfs -m crc=1,finobt=1 -f /dev/mdX where X is the disk #, since I don't use xfs for a long time not current on the options, but probably crc=0 would get rid of the extra space, google "mkfs.xfs man page" for the options. Format will likely not work on a mounted filesystem, you'd need to manually unmount, probably easier for you to do this: -stop array -click on the disk you want to format and change it to a different filesystem -start array, that disk won't mount -type the correct mkfs command for the disk on the console (all data on that disk will be deleted) -stop array, change its fs back to xfs/auto -start array, if you used the correct format options it will now use less space. Sorry to dig up an old post but does "mkfs.xfs -m crc=1,finobt=1 -f /dev/mdX" still work? I used it once before with no problems but went to use it again with the following command : mkfs.xfs -m crc=1,finobt=1 -f /dev/mdg G being the disk # and it came back with an error Error accessing specified device /dev/mdg. Or should I be using the actual driver number as in /dev/md7. Been a long time since I used this command so I can remember. Regards
June 14, 20224 yr Community Expert 1 hour ago, JWMutant said: Or should I be using the actual driver number as in /dev/md7. This
June 14, 20224 yr 14 hours ago, JorgeB said: This Thanks very much, I’ll give it a go when I get home. Cheers
June 15, 20224 yr 21 hours ago, JorgeB said: This Good news is I didn't get any errors this time. So thanks very much for that. The down side is the size used didn't change. 41.9GB for a 8TB drive before and 41.9GB after I followed the instructions.
June 15, 20224 yr Community Expert 1 hour ago, JWMutant said: The down side is the size used didn't change. I didn't noticed before, you quoted the default Unraid command, use the one below that posted by Forusim.
June 15, 20224 yr That worked perfectly. Thanks very much. When I first tried it I thought I may have gotten the command wrong so I did mkfs.xfs -m crc=1,finobt=1 -f /dev/sdg as that is at the end of the description on that drive. It did it thing ok and didnt give me any errors, obviously it didnt work, however what damage if any would I have done by using sdg instead of the correct command? Regards
August 10, 20223 yr On 7/30/2020 at 7:04 AM, JorgeB said: It's normal for newer xfs file systems, they use more space for reflink support and such. Just wanted to confirm, that this explains some pretty huge difference in used space on empty array disks. I emptied 3 identical 10 TB disks on my array and - 2 older ones have ~10GB used space, while the newest one has ~70GB. I assume, the 3rd disk was formatted with the newer version of xfs? Same with the newest 18TB drives, I guess? They all have 126GB used space freshly formatted, instead of ~18GB... Would very much appreciate, if anyone could confirm, that I'm understanding this correctly.
August 11, 20223 yr Community Expert 8 hours ago, shEiD said: Would very much appreciate, if anyone could confirm, that I'm understanding this correctly. You are.
August 11, 20223 yr How can I find out which array disks are actually formatted the "old" and "new" way? I mean, can I get that info without emptying the disks to see how much free space they take up "on empty". Also, what would be the easiest and fastest way to re-format the array disks using the "new" xfs format? I am constantly using hardlinks - I probably have tens of thousands of them on my array, and I am not 100% sure, but it seems, like I am getting way faster results when searching for hardlinks on the "newer" disks, than older ones. or am I just imagining this?
August 12, 20223 yr Community Expert 8 hours ago, shEiD said: How can I find out which array disks are actually formatted the "old" and "new" way? Check the syslog during disk mounting: Quote Aug 11 19:39:28 Kraken root: meta-data=/dev/md17 isize=512 agcount=48, agsize=61047660 blks Aug 11 19:39:28 Kraken root: = sectsz=512 attr=2, projid32bit=1 Aug 11 19:39:28 Kraken root: = crc=1 finobt=1, sparse=1, rmapbt=0 Aug 11 19:39:28 Kraken root: = reflink=1 bigtime=0 inobtcount=0 Aug 11 19:39:28 Kraken root: data = bsize=4096 blocks=2929721331, imaxpct=25 Aug 11 19:39:28 Kraken root: = sunit=0 swidth=0 blks Aug 11 19:39:28 Kraken root: naming =version 2 bsize=4096 ascii-ci=0, ftype=1 Aug 11 19:39:28 Kraken root: log =internal log bsize=4096 blocks=119233, version=2 Aug 11 19:39:28 Kraken root: = sectsz=512 sunit=0 blks, lazy-count=1 Aug 11 19:39:28 Kraken root: realtime =none extsz=4096 blocks=0, rtextents=0 reflink=1 new format, reflink=0 old format. 8 hours ago, shEiD said: Also, what would be the easiest and fastest way to re-format the array disks using the "new" xfs format? There's only one way, empty the disk, reformat, copy the data back.
August 12, 20223 yr 13 hours ago, shEiD said: it seems, like I am getting way faster results when searching for hardlinks on the "newer" disks, than older ones. or am I just imagining this? Speculation here, but it's probably a combination of the new filesystem and the older disks having more fragmentation. Doing a fresh copy to a newly formatted disk solves both issues.
August 16, 20223 yr On 8/12/2022 at 3:50 AM, JorgeB said: There's only one way, empty the disk, reformat, copy the data back. I am using unRAID v6.10.3 My array is data disks only - NO PARITY. What would be the correct procedure to re-format array disks? I would love to be able to do it fast and simple: - stop the array - remove the disk from an array slot (it appears in the Unassigned Devices) - delete the partition with UD - put the disk back in it's array slot - start the array and hopefully the disk gets formatted 🤞 without zeroing out? Or do I need to do it long way round: - new config - remove disk - pre-clear - new config - add disk back - auto-format on array start (the usual unraid behavior when adding a new disk to the array)
August 16, 20223 yr Community Expert 11 minutes ago, shEiD said: What would be the correct procedure to re-format array disks? -stop array, click on the empty disk, change fs to reiserfs -start array, format disk -stop array, click on the disk, change fs back to xfs -start array, format disk
August 16, 20223 yr Community Expert 1 minute ago, shEiD said: What would be the correct procedure to re-format array disks? Same whether you have parity or not. 1 minute ago, JorgeB said: -stop array, click on the empty disk, change fs to reiserfs -start array, format disk -stop array, click on the disk, change fs back to -start array, format disk
August 16, 20223 yr 38 minutes ago, JorgeB said: -stop array, click on the empty disk, change fs to reiserfs -start array, format disk -stop array, click on the disk, change fs back to xfs -start array, format disk Nice and simple - love it. Never used reiserfs before 😀 37 minutes ago, trurl said: Same whether you have parity or not. I thought I'd mention me not using parity, because with parity you would need to zero out the disk?
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.