June 22, 201412 yr BTRFS SUCKs it makes the cache drive really slow reason why i know is when running crashplan on the cache drive it slow down to a crawl ... and crashplan was not quick to begin with ..... top is getting like 20 kworkers running around and every few minutes btrfs messages .... :'( root@R2D2:/# ps x | grep kworker 5 ? S< 0:00 [kworker/0:0H] 13 ? S< 0:00 [kworker/1:0H] 17 ? S< 0:00 [kworker/2:0H] 21 ? S< 0:00 [kworker/3:0H] 1271 ? S 0:09 [kworker/0:1] 3075 ? S 0:05 [kworker/u8:4] 3077 ? S 0:08 [kworker/u8:9] 7115 ? S 0:00 [kworker/3:4] 7343 ? S< 20:04 [kworker/0:1H] 7824 ? S 0:00 [kworker/2:0] 8252 ? S< 12:21 [kworker/2:1H] 8257 ? S< 12:04 [kworker/1:1H] 8258 ? S< 4:02 [kworker/3:1H] 8516 ? S 0:04 [kworker/u8:0] 9122 ? S 0:02 [kworker/u8:3] 10107 ? S 0:02 [kworker/u8:7] 10209 ? S 0:00 [kworker/3:0] 10712 ? S 0:00 [kworker/0:2] 10832 ? S 0:00 [kworker/3:8] 10946 ? S 0:31 [kworker/u8:13] 11269 ? S 0:00 [kworker/1:1] 11307 ? S 0:00 [kworker/u8:8] 12576 pts/0 S+ 0:00 grep kworker 15429 ? S 0:23 [kworker/u8:11] 17295 ? S< 0:00 [kworker/u9:2] 18911 ? S 0:19 [kworker/u8:6] 18945 ? S 0:22 [kworker/u8:10] 19565 ? S 0:01 [kworker/2:1] 23033 ? S< 0:00 [kworker/u9:0] 23590 ? S 0:14 [kworker/u8:2] 23874 ? S 0:13 [kworker/u8:12] 23927 ? S 0:00 [kworker/1:2] 24670 ? D 0:01 [kworker/3:1] 30423 ? S 0:08 [kworker/u8:1] and what is all this shit ? /mnt/user/docker/btrfs/subvolumes/018d01da3eb97eea9cd93c8dd0adf511ab3e346738f5a6f7c813c5db872f4a77/ /mnt/user/docker/btrfs/subvolumes/018d01da3eb97eea9cd93c8dd0adf511ab3e346738f5a6f7c813c5db872f4a77-init/ /mnt/user/docker/btrfs/subvolumes/05c0a7ae5bdeb2af4c7f76bf302275d55f06312a8e43e165fe41493839035a59/ /mnt/user/docker/btrfs/subvolumes/0aa1057cdf960184aa08d0e45e2c2927f2bd9ad7b4d371a4fed4af35915e4c35/ /mnt/user/docker/btrfs/subvolumes/0dc4ba3787df20ea7a02a4574f9d71a8c1da03617694a09465b47cc8b558ad86/ /mnt/user/docker/btrfs/subvolumes/0f8c5bbabce98caea40bdf54a0eb84ab2e1029dcf7f87943675dabc853928c94/ /mnt/user/docker/btrfs/subvolumes/13c1d3856c17852302150deb9f6951eb39d37c7f50aa9a2cf87b64bc59f9dd05/ /mnt/user/docker/btrfs/subvolumes/1908386b990f5bd3138577a443532425c9e28bb993f226561e53d317a55eed1a/ /mnt/user/docker/btrfs/subvolumes/1934846441ad56426a772feb1ed363aeba8916ff829db10d35ce4f98e734816e/ /mnt/user/docker/btrfs/subvolumes/1c3336d191a9a4be86aa6b3b3cc8542631339f80684e552e8f51dd4480c5bae7/ /mnt/user/docker/btrfs/subvolumes/22cf1cfa21846f3d4c89c3f96d8831df144151ab852d60e742fc17478803c23c/ /mnt/user/docker/btrfs/subvolumes/23f361102fae912c44247cd8bcc1a1640292a20d4a4953358b86f74228430711/ /mnt/user/docker/btrfs/subvolumes/248e1024ad119fc564d7e6d54cae236ed8c9e4ff351e6e1a30ff2f810c1ce544/ /mnt/user/docker/btrfs/subvolumes/2f03228c7af08ac38976d75137513a868ee3f298261f71012f17b9f57c8a92d4/ /mnt/user/docker/btrfs/subvolumes/2f4b4d6a4a069aba521ca0067b9263e29d7de84d121ff60ee101242bcc36c13b/ /mnt/user/docker/btrfs/subvolumes/2f8f8e38806f371f82c292eba276c9f1a703f9ef869738d8319e3502ca62e2a2/ /mnt/user/docker/btrfs/subvolumes/3489622e48f9126244c9ce3f86349b96530b62d4cf4e3d60a4d2430df50aa21e/ /mnt/user/docker/btrfs/subvolumes/371b5a8919ea9ff8a7e2ae8728cd7c86ae8a3d68181d65d7ec5275049bf059f6/ /mnt/user/docker/btrfs/subvolumes/43dc9b2ed1acbf0deea5361c8817c4ce81df04b1e454a356fb1765312b28e0e0/ /mnt/user/docker/btrfs/subvolumes/466bb76037ba0ea3bb6feb778dfb7099e23d9dab9726b98578202395232a4954/ /mnt/user/docker/btrfs/subvolumes/47401fd0add90e5944cdd6d1fd433ae469acb607e7f02a36af66da839841d0ba/ can't even delete it ? i had 150 gb free on my cache drive now i have only 85 gb nah think of returning back to a reiserfs drive without docker ....
June 22, 201412 yr BTRFS SUCKs it makes the cache drive really slow reason why i know is when running crashplan on the cache drive it slow down to a crawl ... and crashplan was not quick to begin with ..... top is getting like 20 kworkers running around and every few minutes btrfs messages .... :'( and what is all this shit ? /mnt/user/docker/btrfs/subvolumes/018d01da3eb97eea9cd93c8dd0adf511ab3e346738f5a6f7c813c5db872f4a77/ /mnt/user/docker/btrfs/subvolumes/018d01da3eb97eea9cd93c8dd0adf511ab3e346738f5a6f7c813c5db872f4a77-init/ /mnt/user/docker/btrfs/subvolumes/05c0a7ae5bdeb2af4c7f76bf302275d55f06312a8e43e165fe41493839035a59/ /mnt/user/docker/btrfs/subvolumes/0aa1057cdf960184aa08d0e45e2c2927f2bd9ad7b4d371a4fed4af35915e4c35/ /mnt/user/docker/btrfs/subvolumes/0dc4ba3787df20ea7a02a4574f9d71a8c1da03617694a09465b47cc8b558ad86/ /mnt/user/docker/btrfs/subvolumes/0f8c5bbabce98caea40bdf54a0eb84ab2e1029dcf7f87943675dabc853928c94/ /mnt/user/docker/btrfs/subvolumes/13c1d3856c17852302150deb9f6951eb39d37c7f50aa9a2cf87b64bc59f9dd05/ /mnt/user/docker/btrfs/subvolumes/1908386b990f5bd3138577a443532425c9e28bb993f226561e53d317a55eed1a/ /mnt/user/docker/btrfs/subvolumes/1934846441ad56426a772feb1ed363aeba8916ff829db10d35ce4f98e734816e/ /mnt/user/docker/btrfs/subvolumes/1c3336d191a9a4be86aa6b3b3cc8542631339f80684e552e8f51dd4480c5bae7/ /mnt/user/docker/btrfs/subvolumes/22cf1cfa21846f3d4c89c3f96d8831df144151ab852d60e742fc17478803c23c/ /mnt/user/docker/btrfs/subvolumes/23f361102fae912c44247cd8bcc1a1640292a20d4a4953358b86f74228430711/ /mnt/user/docker/btrfs/subvolumes/248e1024ad119fc564d7e6d54cae236ed8c9e4ff351e6e1a30ff2f810c1ce544/ /mnt/user/docker/btrfs/subvolumes/2f03228c7af08ac38976d75137513a868ee3f298261f71012f17b9f57c8a92d4/ /mnt/user/docker/btrfs/subvolumes/2f4b4d6a4a069aba521ca0067b9263e29d7de84d121ff60ee101242bcc36c13b/ /mnt/user/docker/btrfs/subvolumes/2f8f8e38806f371f82c292eba276c9f1a703f9ef869738d8319e3502ca62e2a2/ /mnt/user/docker/btrfs/subvolumes/3489622e48f9126244c9ce3f86349b96530b62d4cf4e3d60a4d2430df50aa21e/ /mnt/user/docker/btrfs/subvolumes/371b5a8919ea9ff8a7e2ae8728cd7c86ae8a3d68181d65d7ec5275049bf059f6/ /mnt/user/docker/btrfs/subvolumes/43dc9b2ed1acbf0deea5361c8817c4ce81df04b1e454a356fb1765312b28e0e0/ /mnt/user/docker/btrfs/subvolumes/466bb76037ba0ea3bb6feb778dfb7099e23d9dab9726b98578202395232a4954/ /mnt/user/docker/btrfs/subvolumes/47401fd0add90e5944cdd6d1fd433ae469acb607e7f02a36af66da839841d0ba/ can't even delete it ? i had 150 gb free on my cache drive now i have only 85 gb nah think of returning back to a reiserfs drive without docker .... well you are running a beta.. looks like maybe tom needs to tweak btrfs with docker? https://github.com/dotcloud/docker/issues/4100
June 22, 201412 yr New version v3.15 (Jun 2014) pile of send fixes (stability, speed) worker threads now use kernel workqueues
June 24, 201412 yr Tom will you make btrfs a option for the array drives in the future? Thornwood I would like to know this also.
June 24, 201412 yr Anyone else add a pool to the cache drive. It worked until reboot. I followed the instructions to do the RAID 1 pool, got it working, and incorrectly assumed that it would be persistent. Though nothing else with unRAID betas are persistent, so don't know why I assumed. So not knowing anything about BTRFS, I have two questions: [*]How to restart the RAID 1 balance, without losing any data already existing on the drive. [*]How to keep this persistent.
June 24, 201412 yr Anyone else add a pool to the cache drive. It worked until reboot. I followed the instructions to do the RAID 1 pool, got it working, and incorrectly assumed that it would be persistent. Though nothing else with unRAID betas are persistent, so don't know why I assumed. So not knowing anything about BTRFS, I have two questions: [*]How to restart the RAID 1 balance, without losing any data already existing on the drive. [*]How to keep this persistent. I think that was covered in this post. Where you will need to mount it in your go script but not by drive letter since that would/could change but by /dev/by-id.
June 26, 201412 yr Hi All, have had a look around but cannot find a simple answer.... looking at using docker in the near future so know I need to format my cache drive as btrfs, however can I just stop my array, copy everything from the cache drive to another location format to btrfs and then copy all data back? is it really as simple as that? also the actual formatting, am I able to do this from ssh into unRaid? or do I need to remove the drive and format it somewhere else? (I have a windoze box and a mac box) thanks in advance
June 26, 201412 yr Hi All, have had a look around but cannot find a simple answer.... looking at using docker in the near future so know I need to format my cache drive as btrfs, however can I just stop my array, copy everything from the cache drive to another location format to btrfs and then copy all data back? is it really as simple as that? Yes. also the actual formatting, am I able to do this from ssh into unRaid? or do I need to remove the drive and format it somewhere else? (I have a windoze box and a mac box) You can do it from the unRAID GUI. Once the array is stopped, if you click on the cache drive entry then you are taken to a dialog where you can change the format to btrfs. On restarting the cache disk will show as unformatted and there will be an option to format the drive. Do this and the drive is ready for copying back the files. After doing this check that any cache-only shares are set correctly. If you are not careful they may be changed to no longer be cache-only which can have unforeseen side-effects as mover later tries to move files to the main array.
June 26, 201412 yr Hi All, have had a look around but cannot find a simple answer.... looking at using docker in the near future so know I need to format my cache drive as btrfs, however can I just stop my array, copy everything from the cache drive to another location format to btrfs and then copy all data back? is it really as simple as that? also the actual formatting, am I able to do this from ssh into unRaid? or do I need to remove the drive and format it somewhere else? (I have a windoze box and a mac box) thanks in advance Pretty much, see here... http://lime-technology.com/forum/index.php?topic=33806.msg312677#msg312677
June 28, 201412 yr if Plus or Pro users do not wish to reformat their existing reiserfs cache drive device with Btrfs yet can also follow the manual method to test out Btrfs on a different device (non-array / non-cache)... In this Beta, Plus and Pro customers will now be able to format their cache drive feature benefit with the Btrfs file system, directly from within the webGUI. You can do this with either a new or existing device... I booted into b6 and am trying to follow the "webGUI Method" to format my existing cache disk (yeah I already backed up the data) as btrfs. I stopped the array, and it correctly shows the cache disk is already assigned. I can unassign it, then reassign it, but I never get the option to format it. Not sure what I'm missing? Do I have to use the command line method if I want to format an existing reiserfs cache drive as btrfs?
June 28, 201412 yr Not sure what I'm missing? Do I have to use the command line method if I want to format an existing reiserfs cache drive as btrfs? Oh I found it. After stopping the array you have to click on the cache drive to see the details, then you can change the format type to btrfs. Then start the array and format it.
June 30, 201412 yr I need to file a bug / feature / report. I used the command line guide to format a new SSD to btrfs, mounted it, and copied the current contents of my cache drive to it using mc. I then shut down, took the old drive out, and started up in safe mode. When I assigned the drive to the cache slot and started the array, it showed unformatted. When I went into the cache slot properties, the format was still set to reiserfs. I stopped the array, changed it to btrfs, started the array, and presto, there were my files. So, not a real bug, just a request to add the ability to check the format of a drive when it is added to a data or cache slot, and if it contains a valid btrfs or reiserfs partition, apply the appropriate setting automatically.
June 30, 201412 yr Author I need to file a bug / feature / report. I used the command line guide to format a new SSD to btrfs, mounted it, and copied the current contents of my cache drive to it using mc. I then shut down, took the old drive out, and started up in safe mode. When I assigned the drive to the cache slot and started the array, it showed unformatted. When I went into the cache slot properties, the format was still set to reiserfs. I stopped the array, changed it to btrfs, started the array, and presto, there were my files. So, not a real bug, just a request to add the ability to check the format of a drive when it is added to a data or cache slot, and if it contains a valid btrfs or reiserfs partition, apply the appropriate setting automatically. Thank you for letting us know. We saw this happen just recently in our lab as well. Working on a fix...
July 6, 201411 yr Just changed the formatting of my cache drive to btrfs, What can I say, I like new stuff :-) Wondering though.. Is there any sense in having a btrfs cache drive that is used as cachedrive only ? No docker/virtualisation at all ?
July 7, 201411 yr Just changed the formatting of my cache drive to btrfs, What can I say, I like new stuff :-) Wondering though.. Is there any sense in having a btrfs cache drive that is used as cachedrive only ? No docker/virtualisation at all ? Yes if it's sdd. I was wondering if anyone tried to make a pool on the cache drive. Use 2 ssd's for cache drive. I did but not sure what to do on reboot.
July 7, 201411 yr Just changed the formatting of my cache drive to btrfs, What can I say, I like new stuff :-) Wondering though.. Is there any sense in having a btrfs cache drive that is used as cachedrive only ? No docker/virtualisation at all ? Yes if it's sdd. I was wondering if anyone tried to make a pool on the cache drive. Use 2 ssd's for cache drive. I did but not sure what to do on reboot. Its a 256gb ssd.. I do notice that the transferspeeds are quite fast, is that the benefit I should expect ? Or are their tricks under the hood (besides pooling and virtualisation) that I have not yet found ?
July 7, 201411 yr I think one benefit is that those fast speeds will stay fast since it supports trim and reiserfs doesn't.
July 11, 201411 yr Thought I would share my findings with converting my cache drive to btrfs. Cache Drive: 750 GB WD Black 7200 rpm Mounted SSD: 250 GB Samsung SSD Before converting my cache drive was reiserfs and my SSD was ext4. I transferred all of my files from the cache to the SSD with speeds around 80-90 MB/s After converting my cache drive to brtfs and the SSD is still ext4. I transferred all of my files from the SSD to the cache with speeds around 125-130 MB/s and large files would be around 180-200 MB/s I have to say that I am a fan so far.
July 11, 201411 yr That is a lot better then I get... I will be moving unraid out of my vmware environment to a second bare-metal server to see if that will increase speeds..
July 11, 201411 yr Thought I would share my findings with converting my cache drive to btrfs. Cache Drive: 750 GB WD Black 7200 rpm Mounted SSD: 250 GB Samsung SSD Before converting my cache drive was reiserfs and my SSD was ext4. I transferred all of my files from the cache to the SSD with speeds around 80-90 MB/s After converting my cache drive to brtfs and the SSD is still ext4. I transferred all of my files from the SSD to the cache with speeds around 125-130 MB/s and large files would be around 180-200 MB/s I have to say that I am a fan so far. Welcome to a modern day file system. Reiserfs has been dead forever. They even had reiser4 but that was never added to the Linux Kernel. Reiserfs is fast at small files but how many of us primarily are using small files? XFS is even faster than BTRFS when doing single threaded copies like most of us do. When doing several at once, BTRFS is slightly ahead.
July 11, 201411 yr Thought I would share my findings with converting my cache drive to btrfs. Cache Drive: 750 GB WD Black 7200 rpm Mounted SSD: 250 GB Samsung SSD Before converting my cache drive was reiserfs and my SSD was ext4. I transferred all of my files from the cache to the SSD with speeds around 80-90 MB/s After converting my cache drive to brtfs and the SSD is still ext4. I transferred all of my files from the SSD to the cache with speeds around 125-130 MB/s and large files would be around 180-200 MB/s I have to say that I am a fan so far. Welcome to a modern day file system. Reiserfs has been dead forever. They even had reiser4 but that was never added to the Linux Kernel. Reiserfs is fast at small files but how many of us primarily are using small files? XFS is even faster than BTRFS when doing single threaded copies like most of us do. When doing several at once, BTRFS is slightly ahead. YAY modern day file system! If XFS is even faster, I will have to change all my array disks to XFS when it is added. Is there a way to convert a disk from reiserfs to XFS or would I have to copy the contents of the disk to another disk first?
July 11, 201411 yr YAY modern day file system! If XFS is even faster, I will have to change all my array disks to XFS when it is added. Is there a way to convert a disk from reiserfs to XFS or would I have to copy the contents of the disk to another disk first? You will have to copy from disk to disk. I took the drives out of unRAID and did it on another computer. Even on some of my POS slow spinners I was getting 120 Mb/s so it didn't take long. I HIGHLY recommend that you do not use BTRFS on your unRAID drives (cache drive is okay). BTRFS is not Server Grade Quality yet and has way too many outstanding bugs. Plus, they still have a bunch of features they will be introducing which will spawn more bugs.
July 11, 201411 yr You will have to copy from disk to disk. I took the drives out of unRAID and did it on another computer. Even on some of my POS slow spinners I was getting 120 Mb/s so it didn't take long. Hmmm, I'll have to explain to the wife that a new hard drive may be required... I HIGHLY recommend that you do not use BTRFS on your unRAID drives (cache drive is okay). BTRFS is not Server Grade Quality yet and has way too many outstanding bugs. Plus, they still have a bunch of features they will be introducing which will spawn more bugs. Understood, I have been following your other posts on BTRFS as well (well, I read most of what you post). Thanks for responding back and helping out unRAID overall.
Archived
This topic is now archived and is closed to further replies.