May 8, 201016 yr I got a plus registration key and i want to add a cache drive but i cant, it doesn't show up. I am using version 4.4.2 doesn't this support cache drives?
May 8, 201016 yr I got a plus registration key and i want to add a cache drive but i cant, it doesn't show up. I am using version 4.4.2 doesn't this support cache drives? In your 4.4.2 version of unRAID the cache drive was limited to "Pro" versions only. In a later version of unRAID (4.5-beta8), it was also added to the "Plus" version. Release notes can be found here in the wiki: http://lime-technology.com/wiki/index.php?title=Release_Notes (There have been 13 releases of unRAID since 4.4.2, and actually, there have been 14 if you count the current "4.5.3-TEST" release.) In yet a later version of unRAID (4.5.3) , the need for a cache drive for most people was eliminated since write speeds to the array dramatically were improved when a newer version of the Linux kernel was used. (reports from some users are sustained write speeds of 40MB/s over the LAN with NO cache drive) You can update to 4.5.3, but be aware of a bug when initially starting the array. For some users all the disks will fail to mount, and all will show as unformatted in the web-interface. The fix is to simply Stop the array and press "Start" once more. The disks should then mount as expected. This is a dangerous bug, since when a drive is detected as "unformatted" (actually, fails to mount as a reiser file-system) a button is displayed allowing you to format the unformatted disk. This is NOT what you want to press if all your disks show as un-formatted, since it will format them at your request. Formatting all your disks is NOT likely to be one of the first things on your mind... It can bring a grown man to tears... As of a few days ago, a 4.5.3-TEST version of unRAID is available. It has a disabled "Format" button. (It is there but will not function) You probably will not be able to add a cache drive with it, since the cache drive will need to be formatted to use it. You can use it to take advantage of the speed increases in writing to the array. If you are aware of the 4.5.3 bug, and keep from clicking on the "Format" button if any/all your existing drives show up as unformatted, you should be fine. (Just press "Stop" and then "Start" and the drives should then mount as expected and not show as "Unformatted") Joe L.
May 8, 201016 yr Author i cant use 4.5.3 the network speeds drop to 0 and then lose connection to the tower when i use smb from windows. Even thought i have the network problem with 4.5.3 i still tried the test version and i got the bug where it says its unformated i am trying to stop the array and its going in a loop trying to unmount my cache drive and then it finally after 100's of retry to stop the user shares the tower page doesn't load anymore. Hopefully a manual power down doesn't break anything
May 8, 201016 yr tried the test version and i got the bug where it says its unformated i am trying to stop the array and its going in a loop trying to unmount my cache drive and then it finally after 100's of retry to stop the user shares the tower page doesn't load anymore. That is exactly the time you capture the syslog and email it to Limetech along with a detailed description of what you were doing. That is the reason why they posted that version, so we can help them nail those bugs.
May 8, 201016 yr tried the test version and i got the bug where it says its unformated i am trying to stop the array and its going in a loop trying to unmount my cache drive and then it finally after 100's of retry to stop the user shares the tower page doesn't load anymore. That is exactly the time you capture the syslog and email it to Limetech along with a detailed description of what you were doing. That is the reason why they posted that version, so we can help them nail those bugs. If ANY drive is unable to be un-mounted, it is because it is busy. A drive is busy is ANY file on it is open for reading or writing A drive is busy if ANY directory on it is the "current" directory for any process. A drive is busy if ANY directory on it is being used as a mount point for another file-system A drive is busy if ANY file on it is being used as a swap-file. So, in the simplest case, your server was writing to the cache drive, the drive was busy, and therefore could not be un-mounted. Next most likely, you had logged on and changed directory to the cache drive. Since it was then your "current directory" is could not be un-mounted. Next most likely, your "mover" script was moving a file from the cache drive to a data drive, therefore, it could not be un-mounted. Next most likely, you have an add-on process that is running on a file on the cache-drive. I myself had made a similar mistake a few days ago. I had run out of memory while converting some movie files using ffmpeg (an add-on) and handbrake (an add-on) My "emhttp" was killed by the kernel in an attempt to gain some RAM, since it was idle the longest. I started it back up when I was still "CD'd" in the data disk directory. When I attempted to stop the server, it (emhttp) was the reason the drive could not be un-mounted since the data drive was its current directory. Once the series of "Attempting to un-mount" messages start getting written to the syslog they will continue until you either kill the process keeping the disk busy, or you run out of memory (as I did) and emhttp is killed (as mine was) and you lose control from the web. That is what happened to you. (and me) All that in mind. Unless you cleanly stop any processes you add that are accessing the cache drive (or any data drive), then you will see the Un-mounting" messages. That is not a bug, that is a feature. The bug is that the syslog is allowed to fill and use up all RAM. That bug has been there forever, in all releases. The other bug is all the drives failing to mount when you first start the array. That is the one being investigated in 4.5.3-TEST. (It is not yet fixed, just the 4.5.3 "Format" button made non-functional.)
May 9, 201016 yr Author i know that i always shut down sabnznd before i stop array but i cant get the cache drive to unmount, i can even spin down the drive before i do anything and it doesn't wake up so that means im not using it. umount cache works and in unraid the drive doesn't say unmouninng anymore but it still in logs says and it still doesn't stop May 8 17:57:34 Tower emhttp: _shcmd: shcmd (222): exit status: 1 May 8 17:57:34 Tower emhttp: shcmd (223): rmdir /mnt/user >/dev/null 2>&1 May 8 17:57:34 Tower emhttp: _shcmd: shcmd (223): exit status: 1 May 8 17:57:34 Tower emhttp: Retry unmounting user share(s)... May 8 17:57:39 Tower emhttp: shcmd (224): umount /mnt/user >/dev/null 2>&1 May 8 17:57:39 Tower emhttp: _shcmd: shcmd (224): exit status: 1 May 8 17:57:39 Tower emhttp: shcmd (225): rmdir /mnt/user >/dev/null 2>&1 May 8 17:57:39 Tower emhttp: _shcmd: shcmd (225): exit status: 1 May 8 17:57:39 Tower emhttp: Retry unmounting user share(s)... May 8 17:57:44 Tower emhttp: shcmd (226): umount /mnt/user >/dev/null 2>&1 May 8 17:57:44 Tower emhttp: _shcmd: shcmd (226): exit status: 1 May 8 17:57:44 Tower emhttp: shcmd (227): rmdir /mnt/user >/dev/null 2>&1 May 8 17:57:44 Tower emhttp: _shcmd: shcmd (227): exit status: 1 May 8 17:57:44 Tower emhttp: Retry unmounting user share(s)...
May 9, 201016 yr i know that i always shut down sabnznd before i stop array but i cant get the cache drive to unmount, i can even spin down the drive before i do anything and it doesn't wake up so that means im not using it. umount cache works and in unraid the drive doesn't say unmouninng anymore but it still in logs says and it still doesn't stop May 8 17:57:34 Tower emhttp: _shcmd: shcmd (222): exit status: 1 May 8 17:57:34 Tower emhttp: shcmd (223): rmdir /mnt/user >/dev/null 2>&1 May 8 17:57:34 Tower emhttp: _shcmd: shcmd (223): exit status: 1 May 8 17:57:34 Tower emhttp: Retry unmounting user share(s)... May 8 17:57:39 Tower emhttp: shcmd (224): umount /mnt/user >/dev/null 2>&1 May 8 17:57:39 Tower emhttp: _shcmd: shcmd (224): exit status: 1 May 8 17:57:39 Tower emhttp: shcmd (225): rmdir /mnt/user >/dev/null 2>&1 May 8 17:57:39 Tower emhttp: _shcmd: shcmd (225): exit status: 1 May 8 17:57:39 Tower emhttp: Retry unmounting user share(s)... May 8 17:57:44 Tower emhttp: shcmd (226): umount /mnt/user >/dev/null 2>&1 May 8 17:57:44 Tower emhttp: _shcmd: shcmd (226): exit status: 1 May 8 17:57:44 Tower emhttp: shcmd (227): rmdir /mnt/user >/dev/null 2>&1 May 8 17:57:44 Tower emhttp: _shcmd: shcmd (227): exit status: 1 May 8 17:57:44 Tower emhttp: Retry unmounting user share(s)... You should be able to run fuser -cu /mnt/user and see the process ID of the process holding /mnt/user busy. Joe L.
May 9, 201016 yr i know that i always shut down sabnznd before i stop array but i cant get the cache drive to unmount, i can even spin down the drive before i do anything and it doesn't wake up so that means im not using it. umount cache works and in unraid the drive doesn't say unmouninng anymore but it still in logs says and it still doesn't stop May 8 17:57:34 Tower emhttp: _shcmd: shcmd (222): exit status: 1 May 8 17:57:34 Tower emhttp: shcmd (223): rmdir /mnt/user >/dev/null 2>&1 May 8 17:57:34 Tower emhttp: _shcmd: shcmd (223): exit status: 1 May 8 17:57:34 Tower emhttp: Retry unmounting user share(s)... May 8 17:57:39 Tower emhttp: shcmd (224): umount /mnt/user >/dev/null 2>&1 May 8 17:57:39 Tower emhttp: _shcmd: shcmd (224): exit status: 1 May 8 17:57:39 Tower emhttp: shcmd (225): rmdir /mnt/user >/dev/null 2>&1 May 8 17:57:39 Tower emhttp: _shcmd: shcmd (225): exit status: 1 May 8 17:57:39 Tower emhttp: Retry unmounting user share(s)... May 8 17:57:44 Tower emhttp: shcmd (226): umount /mnt/user >/dev/null 2>&1 May 8 17:57:44 Tower emhttp: _shcmd: shcmd (226): exit status: 1 May 8 17:57:44 Tower emhttp: shcmd (227): rmdir /mnt/user >/dev/null 2>&1 May 8 17:57:44 Tower emhttp: _shcmd: shcmd (227): exit status: 1 May 8 17:57:44 Tower emhttp: Retry unmounting user share(s)... Another possibile command to try... Try determining the files/directories being used with: /usr/bin/fuser -mv /mnt/disk* /mnt/user/* once you know the PID (process ID) you can type ps -fp XX (where XX = PID number) to see the full description of the process. Also, just because you can spin down a drive does not mean a file on it cannot be open for reading or writing, but not currently being read or written. It certainly does not mean the disk is not the current working directory for a process. Joe L.
May 9, 201016 yr Author i found a workaround i just removed the cache drive from the disk.cfg and when i reboot my array starts up fine then i just stop and add the cache drive, so far its working fine, not as fast on the network speed as 4.4.2 was but not as bad as it was before i had no cache drive where it would slow down all the way to 0kb/s.
Archived
This topic is now archived and is closed to further replies.