July 30, 201213 yr Edit---Just noticed that I am 3 RC's behind current. I will update to RC6 and report back if good or bad---EndEdit Edit2---I tried RC5 still no joy - might try RC6 next...---EndEdit2 Edit3---I tried RC6 still no joy. Actually quite sad now ---EndEdit3 Edit4---My fault. Forgot to update config on my backup USB licence key. I am the defect. ---EndEdit4 I am copying a new 406MB file onto my unRAID server, and keep getting reports of "No space left on device". As far as I can tell I have plenty of space: Device : Identification : Free Parity : WDC_WD20EARS-00MVWB0_WD-WCAZA5834326 (sdc) 1953514584 : - disk1 : WDC_WD20EARS-00MVWB0_WD-WCAZA5846957 (sdd) 1953514584 : 0 disk2 : WDC_WD20EARS-00MVWB0_WD-WCAZA5834524 (sde) 1953514584 : 336.1 MB disk3 : ST31000333AS_6TE0GH7R (sdf) 976761527 : 41.39 MB disk4 : WDC_WD20EARX-00PASB0_WD-WCAZAE338764 (sdb) 1953514584 : 1.91 TB flash : ASTONE : 995.16 MB The share in question has the following details: Share settings: Name: TV Comments: tv shows with episodic content Allocation method: Most-free Min. -free space: 0 Split level: 99 Included disks: Excluded disks: Share empty: No SMB Security settings: Export: Yes Security: Public The syslog entry of interest is: Jul 30 19:38:28 Tower shfs/user: shfs_write: write: (28) No space left on device From what I can tell I am expecting the tv share to use whatever disks are available and pic the one with the most free space. When I use windows to copy the file the error I get back is "There is not enough space on tv. You need an additional 86.9 MB to copy these files." This would indicate to me that unRaid is attempting to copy the file to disk2. (Which still doesn't dd up properly, but it's as close as I can get with maths...) A similar but different error is occurring when I try to update a big existing file - a 7GB vhd snapshot file: Jul 30 19:15:33 Tower shfs/user: shfs_rename: rename: /mnt/disk1/UserFiles/Users/jason/My Documents/My Virtual Machines/OracleVirtualBox/VBox VMs/Win7/Win7.vdi /mnt/disk1/UserFiles/Users/jason/My Documents/My Virtual Machines/OracleVirtualBox/VBox VMs/Win7/.fuse_hidden0005edbf00000013 (28) No space left on device Jul 30 19:15:34 Tower shfs/user: shfs_write: write: (28) No space left on device Jul 30 19:16:34 Tower last message repeated 4 times I can kind of understand this as it's doing a rename. I assume the OS is trying to create a temporary backup file, then copy up the new file then switching a file system pointer to the new file, then deleting the temporary backup file. Fails if it can't create the new temp file on the same HDD. Kind of makes sense. And I can work around that by deleting it off /mnt/disk1 then re-copying it up from my workstation. But that leads me to the very situation described initially. Even a new file it getting blocked. For some reason unRAID is not seeing the metric bucket load of free space on disk4 - nearly 2TB. The shares for the vdi file is as follows: Share settings: Name: UserFile Comments: My Documents for each user including email and browser settings Allocation method: Most-free Min. -free space: 0 Split level: 99 Included disks: Excluded disks: Share empty: No SMB Security settings: Export: Yes Security: Public I have attached my syslog for some more detail if that will help. I hope this a PEBCAK thing, but I can't figure this one out... Help! Regards, Jaso syslog.txt
July 30, 201213 yr well the update will probably solve it ... but easiest thing to do is just ... stop your array and once stopped .. start again.... this is if you just added that 2 TB drive ? also on the share page how much free space did you see on the tv share ? you added disk 4 to the tv share ?
July 30, 201213 yr Author Did a stop array -> start array as you suggested. Still no joy. AS for the hard drives that are available for the TV share: I have not explicitly added disk4. I have deliberately left all my "included disks" and excluded disks" blank so I don't have to update every share when I add a new drive. I did a "compute" on the shares page and got this for TV: [b]Name Comment Size Free[/b] TV tv shows with episodic content 1.78 TB 122.38 MB disk1 994.47 GB 7.13 MB disk2 557 GB 40.96 KB disk3 173.04 GB 115.22 MB disk4 58.5 GB 1.91 TB I will now try 5.0-rc5 (rc6 looked a bit scary...)
July 30, 201213 yr I think the issue is your min-free-spare setting. Min. -free space: 0 When windows starts to copy the file, it starts at a length of zero bytes and then appends to it gradually growing to its full size. The purpose of the min-space setting is to prevent the use of a drive unless it has over X bytes available. unRAID is apparently attempting to put the new file in an existing directory since it has plenty of space for the "zero byte long" initial file being created. Set the min-free-space parameter to roughly twice the size of the largest file copied. (it is 4k blocks) You probably need something like 256000 for a min-size of 1GB.
July 30, 201213 yr Author @Joe: I tried min-free-space at 256000, then even 512000. Still no joy. Then did the same after stopping/starting the array. Then tried again after a full reboot. @PeterB: I added the split-level of 99 thinking that maybe an empty split level was forcing unRAID to attempt to squish the files into the same folder on the same drive. I couldn't tell what the default was so I picked 99 - which according to the manual "If you set the Split level high, e.g., 999 for a share, then every directory/file created under that share will get placed on a disk according to Allocation method." This is the latest entry in the syslog: Jul 30 22:10:01 Tower shfs/user: shfs_mkdir: assign_disk: TV/3 Parental Guidance/Master Chef Australia All Stars (28) No space left on device Getting desperate now...I even tried copying the files directly to disk4 with the same file structure, but the files did not appear in the tv shares. This has really got me stumped. I have even tried manually adding disk1, disk2, disk3, disk4 to included-discs (still no luck.) One thing I have noticed, after all the shutdowns and reboots (which were done using the standard procedure) the main page is stating that parity was Last checked on Mon Jul 30 21:44:52 2012 EST, finding 2 errors. Could that be causing trouble. Perhaps I should let it run a full parity check (nearly 5 hours...)? Maybe I should bump up to 5.0-rc6-r8168-test? What i don't get is - if I've got the split-level at 99 and the allocation-method as most-free why would it try to put the file (irrespective of it's size) to anywhere but disk4 where there is 1.91TB free??
July 30, 201213 yr Author OK here is the only remaining weird thing: Where has the spare 1.91 TB on disk4 gone? Here are my shares. Compare the free size against my main page (a bit below) User Shares [b]Name Comment Size Free[/b] Backups compute... 458.58 MB Books ebooks (epub, pdf, txt) and audiobooks compute... 458.58 MB FamilyPics Family Pictures - happy snaps etc compute... 458.58 MB Installs Applications installation files, hardware drivers, operating systems iso's etc compute... 458.58 MB Log Logfiles for autobackups compute... 458.58 MB Movies Feature length films compute... 458.58 MB Movies_TODO compute... 458.58 MB Music mp3, flac, ogg, wmv etc compute... 458.58 MB MusicVids Music videos, singles compute... 458.58 MB PcImages Hard drive images(VPC or dd dumps) compute... 458.58 MB Profiles roaming profiles compute... 458.58 MB TV tv shows with episodic content compute... 458.58 MB TV_TODO tv shows that are yet to be processed compute... 458.58 MB UserFiles My Documents for each user including email and browser settings compute... 458.58 MB [b]Device Identification Size Free[/b] parity WDC_WD20EARS-00MVWB0_WD-WCAZA5834326 (sda) 1953514552 2 TB - disk1 WDC_WD20EARS-00MVWB0_WD-WCAZA5846957 (sdb) 1953514552 2 TB 7.13 MB disk2 WDC_WD20EARS-00MVWB0_WD-WCAZA5834524 (sdc) 1953514552 2 TB 336.24 MB disk3 ST31000333AS_6TE0GH7R (sdd) 976761492 1 TB 115.21 MB disk4 WDC_WD20EARX-00PASB0_WD-WCAZAE338764 (sdf) 1953514552 2 TB 1.91 TB flash ASTONE 1.03 GB 994.92 MB It is as if disk4 does not exist! Have I done something extremely stupid and not added disk4 to the array? Is there something else I am supposed to do? I can't figure it out as disk4 has been chugging along in the array just fine. It's been in there for over two weeks now. (We did have a black out the other night that caused an unclean shutdown, but everything came back up AOK (or so I thought...). Uh oh... what if... no way... --- PEBCAK ALERT --- PEBCAK ALERT --- PEBCAK ALERT --- You can all relax. It's me. Not you. We had a blackout a few nights ago. So I prepped my spare licence USB key with clean bzimage and bzroot. What I forgot to do was update the config files. Which should have contained the details of the 2TB HDD I added a few weeks before. Garr I feel like such a dumbass. A few minutes ago I stopped the array, went to the shares tab and saw that I only had disk1, disk2, disk3 included in the user shares. It's taken me a good three hours to figure this out. What he said --> Such a dumbass :-/ Final test. Stopped array. Added disk4 to User Shares. Restarted array. Copied over some TV files. Worked like a charm. --- PEBCAK ALERT --- PEBCAK ALERT --- PEBCAK ALERT --- I think I need to have a little lie-down and relax with some Olympic Games action. Goodnight everybody. :-)
Archived
This topic is now archived and is closed to further replies.