EMKO

i changed min free space for each share from 0 to 20gb because of a no free space error and then i noticed all the problems of slow cache issues are gone.

How do you guys set up your sickbeard/couchpotato to deal with the cache drive flaw. I have mine set up to download and extract on the cache drive that works great but once its done sickbear/couchpotato will move it to the correct share where all the rest of the tv show and movies are problem with this is when these shares are cached it moves the files below 50mb/s. the movie/tv show still ends up on the same drive kinda dumb that it has to be so slow as its now reading from the cache drive and then writing it back on the same drive lol because i am telling it to move to /mnt/user/tv shows/ and that is cached. Couchpotato has setting to look for movies in a directory and then move to another this works great as its instant. if you set "to" to /mnt/cache/Movies/ and "from" where its downloaded on the cache drive the move is instant vs if you told it to move it to /mnt/user/Movies a cached share even thought it ends up in the same place as /mnt/cache/Movies/ it kills the drives performance and takes a while for the move so anyone streaming will get skipping and horrible performance. This fixes Couchpotato as now post processing doesn't fail BUT you have to create the Movie folder on the cache drive everyday or else Couchpotato will fail to move it as it will say that the folder does not exist. For sickbeard i don't know how to make a fix for that as it looks for tv shows in only one place and moves to the same place. how do you guys set yours up? thanks

The disk is thrashing. This is a relink and if the user file system were a bit smarter it would relink instead of copying. This occurs in systems very low power CPUs, e.g., Atom or AMD Turion. I see whats happening so i have 100mb/s write speed and 100mb/s read when the file gets moved on the same drive its reading and writing at the same time so it drops to around 50mb/s read and write, this causes everything on the cache drive if accessed by smb to be not streamable as the disk is maxed out during the file move. Array will be faster for almost everything i don't see the point of Cache drive now for smb transfers CPU usage is about the same its a AMD dual core tested a few large files caused it to go to around 60-70% CPU one is going at 40-50mb/s, cache disk share is going at 90-105mb/s

i dont see how customization will fix this try it out your self mv some files from cache drive to a /mnt/user/ share that is cached you will see it get moved to the cache drive but at a slow speed why? when using mv to move file from cache drive to the same location the cached share will move it to is instant. even worse transferring from another computer over smb to a cached share is slower then transferring to the cache drive why? really cache is useless as it is now because i can just use the array and it would be faster but i would have to keep waking it up.

can you fix some how the cache shares work? its very slow when you transfer to cached shares vs directly to cache drive share, same with doing any moving/copying on the system its self. example you can test your self file1 is in /mnt/cache/download - this is the cache drive mv file1 /mnt/user/Tv\ Shows/show/ = very slow around 40-50MB/s i think - Tv shows is a cached share so the file even though its going to user/tv shows it ends up on /mnt/cache/Tv Shows/Show/ which is correct but the slow speed is very bad it affects everything on the cache drive. but when your in /mnt/cache/download and do this mv file1 /mnt/cache/Tv\ Shows/show/ = instant is there any way to make unraid understand where the file should be moving/copy to is the cache drive it self and instead make it a normal move/copy? maybe there is a way to make a another cache mnt that includes everything thats on user so applications can see all the files and still be working off the cache device would make it not slow down for that but only the SMB cached shares i guess.

i still cant figure out a way to fix this anyone have any ideas or is this a limitation of unraid.

no i have not had any luck in fact Couchpotato fix i thought i had is not possible since when the Movies folder in cache drive gets moved during the night and when a movie gets downloaded it complains now about ERROR[tato.core.plugins.renamer] Both the "To" and "From" folder have to exist. it has something to do with unraid and how cache shares behave. does anyone have any ideas?

i noticed my windows file history backups has over 400k files and lots of folders so i added -a '-noleaf \( -name Data -prune -o -name data -prune \) -o -print' the folder is /mnt/user/backup/username/folder/Data since i did this i see its running but any time i browse any cached dirs it still wakes up all the drives and takes time so i think its not working correct. Am i using the commands in a wrong way? is there a way to see the status of what is has cached or what it is doing? Jun 12 11:59:59 Tower cache_dirs: ============================================== Jun 12 11:59:59 Tower cache_dirs: command-args=-w -B -a -noleaf ( -name Data -prune -o -name data -prune ) -o -print Jun 12 11:59:59 Tower cache_dirs: vfs_cache_pressure=10 Jun 12 11:59:59 Tower cache_dirs: max_seconds=10, min_seconds=1 Jun 12 11:59:59 Tower cache_dirs: max_depth=9999 Jun 12 11:59:59 Tower cache_dirs: command=find -noleaf ( -name Data -prune -o -name data -prune ) -o -print Jun 12 11:59:59 Tower cache_dirs: version=1.6.7

no that wont work that's just the download dir anyway's sickbeard will move and rename to correct location and i want to use post processing script. sickbeard would have to rename and move the file to /mnt/cache/tv shows/tvshowname and look for the tv shows in /mnt/user/tv shows/tvshowname but sickbeard only moves and looks in the same location so its not possible like it was with Couchpotato maybe i will just download to non cached share, download speed is not greater then array write speed only thing it would effect is extracting speed i think ahh nvm that would still be probably same

does cache read speed get graphed in Dynamix stats page? i see the writes get graphed but not the reads is that normal?

it is writing to cache drive with shares that are set to be cached just very slow array speed compared to directly to cache share disk for the moving of files on cache drive i think Unraid just doesn't know to just move the file and instead is actually reading and then making the file on the same disk so maybe that's why it hits a 50MB/s limit it really should be instant if Unraid knew somehow to just ignore the /mnt/user/share* when its cached and instead just do a mv on the file to the cache dir. i fixed Couchpotato by making it rename files to /mnt/cache/Movies works instant now but i cant figure out a way to do that with sickbeard

so no one has this problem? is it just me? i am running cache_dirs and i have 6GB of ram 330mb used 4.6gb cached is that ok? it got so bad today i would transfer at 16MB/s i don't get it and i did a hdparm -tT /dev/sdk and it showed the same result after a reboot speed came back to 40MB/s on cached shares and 75MB/s+ on cache drive share. i am fine with slower transfer speed for cached dirs but not with slow moving files on the same drive. for example a 8gb file i have to wait for the application to move it more then 3minutes when it should move it instant all because its going to /mnt/user/Movies from /mnt/cache/download even thought /mnt/user/Movies is still moving to /mnt/cache/Movies since its cached. I don't remember this happing before i think i would of noticed maybe its 5.0.5 i don't know. did some tests in console show.mkv is in /mnt/cache/download mv show.mkv /mnt/user/Tv\ Shows/show/ = very slow around 40-50MB/s i think mv show.mkv /mnt/cache/Tv\ Shows/show/ = instant

no and i just did test it does not make a difference, when moving a file in unraid on the cache drive with windows its instant but when unraid application does it like couchpotato it goes the speed of the array drives. for example i can move files on cache drive like file1 to folder1 and its instant on windows with smb share when couchpotato moves file on cache drive from 1 folder and transfers it to say /mnt/user/movies which is still moving it on the same drive it goes the speed of the array if i transfer with smb to /mnt/user/movies it goes the speed of the array share movies is cached share so all files appear on the cache drive when i transfer them so they should not slow down to array speed correct?

/tower/cache i get full speed of the hard drive but when i use any shares that are cached it slows down to array speed as if its transferring to non cached drive. /tower/cache i get 60+MB/s /tower/movies i get 25-40MB/s the file is transferring to /tower/cache/movies that speed is the same as non cached shares. Even when a program downloads a file to the cache drive and then moves it to a cached share it goes very slow lower then 50MB/s and the file is on the Cache drive moving to the Cache drive this should be very fast no? is this normal?

okay it finally finished pre clearing disc5 some things that changed Offline_Uncorrectable used to be 1 and is 0 now Raw_Read_Error_Rate was 33 is now 46 here is the result from pre clear Final SMART report is that what is causing the errors to show next to the drive? Raw_Read_Error_Rate and Seek_Error_Rate and how come Unraid doesn't care it cant read a sector? Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013640 or is it trying to read it fails and then tries again and succeed so it shows up as a error but it doesn't red ball it>? what do you think is the drive okay to use?

yea that's weird but from my drive it looks like it red balled because of the UDMA_CRC_Error_Count something to do with the cables i re pluged it took out disk 5 and i rebuild the data, right now i am waiting for parity verify then i will preclear disk5 to see why its making those errors.

its this drive that red balled http://lime-technology.com/forum/index.php?topic=33070.0 i pre cleared it and looks ok i think i will remove disc5 and put this in and let it rebuild and do a parity test then pre clear disk5 and see if its okay. Is that a good idea? this is the 3tb drive from that link ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x002f 200 200 051 Pre-fail Always - 1 3 Spin_Up_Time 0x0027 176 152 021 Pre-fail Always - 8183 4 Start_Stop_Count 0x0032 097 097 000 Old_age Always - 3284 5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0 7 Seek_Error_Rate 0x002e 100 253 000 Old_age Always - 0 9 Power_On_Hours 0x0032 080 080 000 Old_age Always - 14785 10 Spin_Retry_Count 0x0032 100 100 000 Old_age Always - 0 11 Calibration_Retry_Count 0x0032 100 253 000 Old_age Always - 0 12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 42 192 Power-Off_Retract_Count 0x0032 200 200 000 Old_age Always - 27 193 Load_Cycle_Count 0x0032 187 187 000 Old_age Always - 40146 194 Temperature_Celsius 0x0022 121 109 000 Old_age Always - 31 196 Reallocated_Event_Count 0x0032 200 200 000 Old_age Always - 0 197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 0 198 Offline_Uncorrectable 0x0030 200 200 000 Old_age Offline - 0 199 UDMA_CRC_Error_Count 0x0032 200 200 000 Old_age Always - 6 200 Multi_Zone_Error_Rate 0x0008 200 200 000 Old_age Offline - 1 == WDC WD30EZRX-00MMMB0 WD-WCAWZ1857328 == Disk /dev/sdf has been successfully precleared == with a starting sector of 1 == Ran 1 cycle == == Using :Read block size = 8225280 Bytes == Last Cycle's Pre Read Time : 9:42:49 (85 MB/s) == Last Cycle's Zeroing time : 8:28:48 (98 MB/s) == Last Cycle's Post Read Time : 24:06:49 (34 MB/s) == Last Cycle's Total Time : 42:19:33 == == Total Elapsed Time 42:19:33 == == Disk Start Temperature: 22C == == Current Disk Temperature: 31C, == ============================================================================ ** Changed attributes in files: /tmp/smart_start_sdf /tmp/smart_finish_sdf ATTRIBUTE NEW_VAL OLD_VAL FAILURE_THRESHOLD STATUS RAW_VALUE Spin_Up_Time = 176 155 21 ok 8183 Seek_Error_Rate = 100 200 0 ok 0 Temperature_Celsius = 121 129 0 ok 31 No SMART attributes are FAILING_NOW 1 sector was pending re-allocation before the start of the preclear. 1 sector was pending re-allocation after pre-read in cycle 1 of 1. 0 sectors were pending re-allocation after zero of disk in cycle 1 of 1. 0 sectors are pending re-allocation at the end of the preclear, a change of -1 in the number of sectors pending re-allocation. 0 sectors had been re-allocated before the start of the preclear. 0 sectors are re-allocated at the end of the preclear, the number of sectors re-allocated did not change. ============================================================================

i checked the logs folder but its not there and i have since rebooted the system and the logs are new now and the old one didn't save. I can do another parity check if that will help get the logs again. I have that one 3tb drive that red balled but looks ok maybe i replace this drive with that and do a preclear on it?

yes i can see the file uploading into the cache dir under the correct share folder that will be moved later. like uploading to movies share the file goes to /cache/movies/file at 60MB/s uploading directly to /cache at 105MB/s Now i notice why extracting with sabnzbd was slower on my cache drive I set downloads share to be cached and when i set sabnzbd to extract here its slower then extracting directly to cache if its set to /mnt/user/Downloads/ its maxed out at 60MB/s rar extracting vs /mnt/cache

Last checked on Sat Apr 26 11:19:58 2014 MDT (today), finding 0 errors. ? Duration: 16 hours, 13 minutes, 29 seconds. Average speed: 68.5 MB/sec On the Hard drive list Disc 5 has 22 errors but Parity check found 0 errors tail -n 40 -f /var/log/syslog Apr 26 08:03:53 Tower kernel: sd 2:0:0:0: [sdb] Apr 26 08:03:53 Tower kernel: Result: hostbyte=0x00 driverbyte=0x08 Apr 26 08:03:53 Tower kernel: sd 2:0:0:0: [sdb] Apr 26 08:03:53 Tower kernel: Sense Key : 0x3 [current] [descriptor] Apr 26 08:03:53 Tower kernel: Descriptor sense data with sense descriptors (in hex): Apr 26 08:03:53 Tower kernel: 72 03 11 04 00 00 00 0c 00 0a 80 00 00 00 00 01 Apr 26 08:03:53 Tower kernel: 5b 41 ea 08 Apr 26 08:03:53 Tower kernel: sd 2:0:0:0: [sdb] Apr 26 08:03:53 Tower kernel: ASC=0x11 ASCQ=0x4 Apr 26 08:03:53 Tower kernel: sd 2:0:0:0: [sdb] CDB: Apr 26 08:03:53 Tower kernel: cdb[0]=0x88: 88 00 00 00 00 01 5b 41 e7 50 00 00 03 68 00 00 Apr 26 08:03:53 Tower kernel: end_request: I/O error, dev sdb, sector 5826013704 Apr 26 08:03:53 Tower kernel: ata2: EH complete Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013640 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013648 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013656 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013664 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013672 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013680 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013688 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013696 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013704 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013712 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013720 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013728 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013736 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013744 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013752 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013760 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013768 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013776 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013784 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013792 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013800 Apr 26 08:03:53 Tower kernel: md: disk5 read error, sector=5826013808 Apr 26 08:53:56 Tower kernel: mdcmd (48): spindown 5 Apr 26 11:19:58 Tower kernel: md: sync done. time=58409sec Apr 26 11:19:58 Tower kernel: md: recovery thread sync completion status: 0 Apr 26 12:05:00 Tower kernel: mdcmd (49): spindown 0 Apr 26 12:05:00 Tower kernel: mdcmd (50): spindown 4 ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x002f 200 200 051 Pre-fail Always - 33 3 Spin_Up_Time 0x0027 178 178 021 Pre-fail Always - 6066 4 Start_Stop_Count 0x0032 100 100 000 Old_age Always - 436 5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0 7 Seek_Error_Rate 0x002e 200 200 000 Old_age Always - 0 9 Power_On_Hours 0x0032 083 083 000 Old_age Always - 13107 10 Spin_Retry_Count 0x0032 100 100 000 Old_age Always - 0 11 Calibration_Retry_Count 0x0032 100 253 000 Old_age Always - 0 12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 45 192 Power-Off_Retract_Count 0x0032 200 200 000 Old_age Always - 30 193 Load_Cycle_Count 0x0032 166 166 000 Old_age Always - 103526 194 Temperature_Celsius 0x0022 133 119 000 Old_age Always - 17 196 Reallocated_Event_Count 0x0032 200 200 000 Old_age Always - 0 197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 0 198 Offline_Uncorrectable 0x0030 200 200 000 Old_age Offline - 1 199 UDMA_CRC_Error_Count 0x0032 200 200 000 Old_age Always - 0 200 Multi_Zone_Error_Rate 0x0008 200 200 000 Old_age Offline - 12 What do i do? is the drive bad? thanks

Any ideas if this can be fixed? i get 105MB/s on cache but on cached shares i get same speed as non cached shares 60MB/s

Could UDMA_CRC_Error_Count cause the red ball? is this where cable could be the problem? does the drive look okay to add back to the array? Before the test there was 1 current_pending_sector ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x002f 200 200 051 Pre-fail Always - 1 3 Spin_Up_Time 0x0027 176 152 021 Pre-fail Always - 8183 4 Start_Stop_Count 0x0032 097 097 000 Old_age Always - 3284 5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0 7 Seek_Error_Rate 0x002e 100 253 000 Old_age Always - 0 9 Power_On_Hours 0x0032 080 080 000 Old_age Always - 14785 10 Spin_Retry_Count 0x0032 100 100 000 Old_age Always - 0 11 Calibration_Retry_Count 0x0032 100 253 000 Old_age Always - 0 12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 42 192 Power-Off_Retract_Count 0x0032 200 200 000 Old_age Always - 27 193 Load_Cycle_Count 0x0032 187 187 000 Old_age Always - 40146 194 Temperature_Celsius 0x0022 121 109 000 Old_age Always - 31 196 Reallocated_Event_Count 0x0032 200 200 000 Old_age Always - 0 197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 0 198 Offline_Uncorrectable 0x0030 200 200 000 Old_age Offline - 0 199 UDMA_CRC_Error_Count 0x0032 200 200 000 Old_age Always - 6 200 Multi_Zone_Error_Rate 0x0008 200 200 000 Old_age Offline - 1 == WDC WD30EZRX-00MMMB0 WD-WCAWZ1857328 == Disk /dev/sdf has been successfully precleared == with a starting sector of 1 == Ran 1 cycle == == Using :Read block size = 8225280 Bytes == Last Cycle's Pre Read Time : 9:42:49 (85 MB/s) == Last Cycle's Zeroing time : 8:28:48 (98 MB/s) == Last Cycle's Post Read Time : 24:06:49 (34 MB/s) == Last Cycle's Total Time : 42:19:33 == == Total Elapsed Time 42:19:33 == == Disk Start Temperature: 22C == == Current Disk Temperature: 31C, == ============================================================================ ** Changed attributes in files: /tmp/smart_start_sdf /tmp/smart_finish_sdf ATTRIBUTE NEW_VAL OLD_VAL FAILURE_THRESHOLD STATUS RAW_VALUE Spin_Up_Time = 176 155 21 ok 8183 Seek_Error_Rate = 100 200 0 ok 0 Temperature_Celsius = 121 129 0 ok 31 No SMART attributes are FAILING_NOW 1 sector was pending re-allocation before the start of the preclear. 1 sector was pending re-allocation after pre-read in cycle 1 of 1. 0 sectors were pending re-allocation after zero of disk in cycle 1 of 1. 0 sectors are pending re-allocation at the end of the preclear, a change of -1 in the number of sectors pending re-allocation. 0 sectors had been re-allocated before the start of the preclear. 0 sectors are re-allocated at the end of the preclear, the number of sectors re-allocated did not change. ============================================================================

I notice one of my drives is always awake Windows file history keeps check the configuration folder on my array every 5 minutes to see if it needs to back up anything i think. Is there anyway to have this be done on the cache drive? so far it does backup my files to the cache drive and then the files get moved over night its just the configuration folder that it keeps reading that wakes up the drive.

I replaced a 3tb parity with a 4tb that worked fine then i took the 3tb and put it in the server as disc 6 when i tried to clear it with preclear it caused Disc 4 and Cache to mess up so when unraid failed to write it set DISC 4 as RED BALL. I tested if it was the SASLP-MVL8 by replacing it with the other one i have and that didn't work, i swapped one of the cables so it booted on slot 7 instead of the slot 3 cable and that still didnt work so now i cant tell is the hotswap bay broken is it the power or sata connection. If its the power connection or sata how does this cause a problems for the other drives on the controller? Right now its working in the bay next to it on the same controller and same power as the one next to it. I put the 3tb in the place of DISC 4 to rebuild the data next i will try to pre clear disc 4 and see if it is just that one bay. the case is a norco 4020 my syslog is here if anyone thinks it might help http://lime-technology.com/forum/index.php?topic=28382.msg253823#msg253823 Thanks

Just took the server apart all drives that failed once i added this drive are all on the same controller Supermicro AOC-SASLP-MVL8 DISC 4 red ball and cache failed while trying to clear DISC 6 witch is also on the same controller how do i tell unraid that this hard drive is fine so i don't have to rebuild it? and whats causing this problem? the controller?