Ruthalas Posted April 5, 2018 Share Posted April 5, 2018 (edited) How is a cache pool of disparately-sized drives handled? I currently have a 60GB SSD and a 120GB SSD pooled for my cache. The result is listed at 90GB of cache space. That seems wrong for either RAID 0 or RAID 1 in my limited understanding.This post seems like it details alternate options for this configuration, but doesn't shed light of the default behavior. (I was considering swapping the 120GB drive out for a matching 60GB drive to use the larger drive elsewhere, but I want to understand exactly how the current system is working before I make that decision.) Edited April 5, 2018 by Ruthalas Quote Link to comment
JorgeB Posted April 5, 2018 Share Posted April 5, 2018 http://carfax.org.uk/btrfs-usage/ 1 Quote Link to comment
John_M Posted April 5, 2018 Share Posted April 5, 2018 6 hours ago, johnnie.black said: http://carfax.org.uk/btrfs-usage/ And, according to that calculator, the RAID 0 capacity is 120 GB and the RAID 1 capacity is 60 GB, as you might expect. So why is the OP seeing 90 GB reported? Is it a BTRFS bug or an unRAID bug? I vaguely remember this cropping up before and I'm surprised it hasn't been fixed yet. @Ruthalas Would you mind posting your diagnostics so we can see what BTRFS thinks is going on? 1 Quote Link to comment
JorgeB Posted April 5, 2018 Share Posted April 5, 2018 2 minutes ago, John_M said: So why is the OP seeing 90 GB reported? Is it a BTRFS bug or an unRAID bug? I vaguely remember this cropping up before and I'm surprised it hasn't been fixed yet. It's a btrfs limitation: https://lime-technology.com/forums/topic/46802-faq-for-unraid-v6/?do=findComment&comment=480420 2 Quote Link to comment
Ruthalas Posted April 5, 2018 Author Share Posted April 5, 2018 For my understanding- that means I have 60GB of usable space, regardless of the reported size, yes? In that case I should swap to the matching-size drive, as no benefit in conferred by having one larger drive. How does the process of changing a cache disk out differ when you have 2 disks, if at all? Quote Link to comment
JorgeB Posted April 5, 2018 Share Posted April 5, 2018 1 minute ago, Ruthalas said: For my understanding- that means I have 60GB of usable space, regardless of the reported size, yes? Correct. 2 minutes ago, Ruthalas said: How does the process of changing a cache disk out differ when you have 2 disks, if at all? You can use this one instead, much simpler and quicker for a pool. 1 Quote Link to comment
Ruthalas Posted April 5, 2018 Author Share Posted April 5, 2018 Ah, excellent. That's perfect. Thank you! Quote Link to comment
Ruthalas Posted April 6, 2018 Author Share Posted April 6, 2018 (edited) I am having an issue, I believe due to the fact that I am attempting to move to a smaller drive. I receive this error: BTRFS: error (device sdj1) in btrfs_create_pending_block_groups :10269: errno=-28 No space left Any advice on how to approach that? Edit: I have backed up the full contents of the cache pool. Edited April 6, 2018 by Ruthalas Quote Link to comment
JorgeB Posted April 6, 2018 Share Posted April 6, 2018 1 hour ago, Ruthalas said: I believe due to the fact that I am attempting to move to a smaller drive. You can't replace a device with a smaller one using that procedure, it warns so in the notes. Quote Link to comment
Ruthalas Posted April 6, 2018 Author Share Posted April 6, 2018 Can I rebuild the entire pool, and then copy the contents back on? Quote Link to comment
JorgeB Posted April 6, 2018 Share Posted April 6, 2018 If the pool still mounts use this procedure to backup, rebuild pool, restore, if it doesn't post current diags. Quote Link to comment
Ruthalas Posted April 6, 2018 Author Share Posted April 6, 2018 I am attempting to follow those instructions but am running into one issue- The system share won't move. I have set it to 'Yes', but a single file (system\libvirt\libvirt.img) won't move when I invoke the mover. Am I missing something, or is there a way to force that move? Quote Link to comment
JorgeB Posted April 6, 2018 Share Posted April 6, 2018 The obvious reason would be the file being in use, did you shutdown the VM service? Quote Link to comment
Ruthalas Posted April 6, 2018 Author Share Posted April 6, 2018 (edited) I did: Under Settings>Vm Manager>Enable VMs is set to "No" Is there something else that might be using that? Edit: (And if I didn't follow the proper method to disable VMs, let me know!) Edited April 6, 2018 by Ruthalas Quote Link to comment
JorgeB Posted April 6, 2018 Share Posted April 6, 2018 Probably not shutting down correctly, leave it disable, reboot and run the mover. 1 Quote Link to comment
Ruthalas Posted April 6, 2018 Author Share Posted April 6, 2018 (edited) A restart did not allow that file to be moved. (afterwards) I toggled the VM enable back on and off as well, with no success. Would I be able to see errors that it may be throwing in the diagnostics download? Edited April 6, 2018 by Ruthalas Quote Link to comment
Ruthalas Posted April 6, 2018 Author Share Posted April 6, 2018 Can I simply back that file up, perform the swap, and restore the file? Quote Link to comment
JorgeB Posted April 7, 2018 Share Posted April 7, 2018 Can I simply back that file up, perform the swap, and restore the file? Yes, if successfully copied, there's a chance it's not being moved because it's corrupt, diagnostics would show. 1 Quote Link to comment
Ruthalas Posted April 8, 2018 Author Share Posted April 8, 2018 I was able to ignore that single file and copy it over manually after a rebuild. Thanks for your help man! Quote Link to comment
Recommended Posts
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.