I have mine and my wife's music collection on my unraid server.  It used to be that when itunes needed a file that was on unraid, it would spin up the drives that were needed, and everything worked fine.  My wife would complain about how that first file could take a while to open, but otherwise it was all good.  Now for some reason even if the drives are spun up, itunes can't find the music file unless you actually go to the mapped drive under my computer and open it.  Then itunes finds the files just fine.  Keep in mind that I'm not losing the drive mapping.  That is in place and fine.  It's like for some reason itunes doesn't want to use it until something else on the computer does.  It's weird.  Any ideas?

do you need the drives to spin down or can you run them without doing so? It won't hurt the drives, they are built for running flat out all the time unless power/heat is a concern.

 

I've turned off spin down on my drives for this reason, no point really, let those babies go!

 

Not that anyone cares, but going back to 4.3.3 fixed this problem.  I went down this road after reading about some slowness issues others were having, that I had noticed as well.  Oh well, lets hope 4.4.3 is a winner. 

Make sure to report this to Tom.

 

I just went into iTunes, drive containing my mp3s was spun down, and played a song.  Took a couple extra seconds but started playing just fine.

 

Not sure what you are seeing, but may have something to do with your configuration.

 

(I am not running the latest iTunes BTW.  Not sure if that matters.)

it could matter because itunes is known to change core characteristics between version (doesn't have to be a major version change)

 

 

One more data point:  I just reread the OP.  He is using mapped drives.  I am not.  I am just using the "//tower/mps3" network path.

I'm also having errors using unraid as a datastore with my mac (for iTunes library, iPhoto library, etc)

 

I have unraid pro on a PC, with a user share shared via gige. My client is an iMac running the latest leopard os x 10.5 (but the problem has been here since I started using unraid w/earlier 10.5's). I have a mapped volume at /Volumes/files that is mapped via the directory utility to nfs connect to my unraid server at files:/mnt/bitpool ... I tried a basic, no-frills mount at first, got the errors, and tried many variations (currently rsize=32768 wsize=32768 intr noatime). On unraid, the user share has "*(rw,insecure)" for nfs, and I've tried with and without a cache disk.

 

If I buy an album off iTunes (with my local ~/Music/iTunes Library a symlink to a dir on unraid), as the album downloads, the first file succeeds, followed by the second file barfing after a bit with an error -36 or something, followed by a success, barf, success, barf, etc.. Meanwhile, sometimes my library will become corrupted, and it can no longer use the iTunes Library file (I have to readd my iTunes Music folder). I have similar problems with iPhoto - it will sometimes complain that it can't write to the iPhoto library. If I copy files, etc., from the command line, I never seem to see any issues. Any ideas, solutions? Maybe use a different mounting method? In macosx, you can use directory utility to automount, use fstab, mount via the finder, and finally just access /net/files/mnt/user/bitpool ....

 

This problem has led me to somewhat regret my purchase of two pro keys and wish I had looked into installing freebsd w/ZFS. But there you have to choose between a mix/match drive size expandable pool with best-effort (in terms of spreading across devices) file-level replication (using the "copies" attribute - but cool that it can be ==2 for some directories, 1 for others, 3 for others), or adding 2 mirrored same-size drives at a time, or adding 3 same-size drives in 2/3 efficient parity config... I would probably go for the big pool with file-level copies.

Any update on this?

 

I am getting weird errors accessing unRAID from a Mac after everything sits unused for a while.

 

I have spindown set to 1-hour and after a drive spins down, I can go into Finder on the mac, access something on that drive, and I can see after a pause, it spins up, servers the files just fine. However, if I've been away for a while longer (don't know exactly how long that is - maybe several hours), when I come back it won't spin up automatically. I can go manually spin them up but it still doesn't matter. I "eject" all my shares and try to re-mount them, but I get "error -36". So far the only solution I've found is to reboot the unRAID server. Then everything works great until the next time.

 

I know that description sounds like a "user" description, but at this point, I can't find anything that is common between times when this happens. I don't loose the shares - they still show mounted. One thing's for sure, when it does happen, I haven't found anything that'll get me access except for rebooting unRAID.

 

My next test is to set spindown to "never" and see if it goes away (but I really don't think it will). Just rule out one thing at a time and then maybe I can find a clue. Also, I have another few Macx, but I haven't tried to access the shares from them while access from another one is fouled up - I'll test that too.

 

FYI: I'm sharing everything via samba, ethernet is hard-wired, the Mac I'm using never goes to sleep...can't think of anything else helpful - ask me if something comes to mind.

 

Thanks

 

I had forgotten about this post.  I'm pretty sure it was an Itunes or a Windows problem of some kind.  Not a problem with Unraid.  Not long after I posted this, maybe a week, my wifes computer would no longer boot into Windows, and I could not figure out why.  So, I ended up re-installing, with XP sp3 slipstreamed.  Of course machine ended up with the latest itunes, whatever it was at that time. 

 

A week or so later, just for the hell of it, I put 4.4.2 back on the unraid, and now it works as expected.  Drives spin up when Itunes attempts access and aside from initial slowness, which she still bitches about, everything is good. 

 

I'm clueless on macs, so sorry I can't help you OSX folks. 

I'm also having errors using unraid as a datastore with my mac (for iTunes library, iPhoto library, etc)

 

I have unraid pro on a PC, with a user share shared via gige. My client is an iMac running the latest leopard os x 10.5 (but the problem has been here since I started using unraid w/earlier 10.5's). I have a mapped volume at /Volumes/files that is mapped via the directory utility to nfs connect to my unraid server at files:/mnt/bitpool ... I tried a basic, no-frills mount at first, got the errors, and tried many variations (currently rsize=32768 wsize=32768 intr noatime). On unraid, the user share has "*(rw,insecure)" for nfs, and I've tried with and without a cache disk.

 

If I buy an album off iTunes (with my local ~/Music/iTunes Library a symlink to a dir on unraid), as the album downloads, the first file succeeds, followed by the second file barfing after a bit with an error -36 or something, followed by a success, barf, success, barf, etc.. Meanwhile, sometimes my library will become corrupted, and it can no longer use the iTunes Library file (I have to readd my iTunes Music folder). I have similar problems with iPhoto - it will sometimes complain that it can't write to the iPhoto library. If I copy files, etc., from the command line, I never seem to see any issues. Any ideas, solutions? Maybe use a different mounting method? In macosx, you can use directory utility to automount, use fstab, mount via the finder, and finally just access /net/files/mnt/user/bitpool ....

 

This problem has led me to somewhat regret my purchase of two pro keys and wish I had looked into installing freebsd w/ZFS. But there you have to choose between a mix/match drive size expandable pool with best-effort (in terms of spreading across devices) file-level replication (using the "copies" attribute - but cool that it can be ==2 for some directories, 1 for others, 3 for others), or adding 2 mirrored same-size drives at a time, or adding 3 same-size drives in 2/3 efficient parity config... I would probably go for the big pool with file-level copies.

 

This is pretty old, hope you have received some help, but if not:

 

It sounds like your issues are with the new and still very beta support of NFS, so I would email Tom directly at [email protected].  He has indicated he needed feedback on the new NFS support, but I've seen very little, so it does not seem as if many are using it.  I'm sure he would value your cooperation in fixing this problem.

Any update on this?

 

I am getting weird errors accessing unRAID from a Mac after everything sits unused for a while.

 

I have spindown set to 1-hour and after a drive spins down, I can go into Finder on the mac, access something on that drive, and I can see after a pause, it spins up, servers the files just fine. However, if I've been away for a while longer (don't know exactly how long that is - maybe several hours), when I come back it won't spin up automatically. I can go manually spin them up but it still doesn't matter. I "eject" all my shares and try to re-mount them, but I get "error -36". So far the only solution I've found is to reboot the unRAID server. Then everything works great until the next time.

 

I know that description sounds like a "user" description, but at this point, I can't find anything that is common between times when this happens. I don't loose the shares - they still show mounted. One thing's for sure, when it does happen, I haven't found anything that'll get me access except for rebooting unRAID.

 

My next test is to set spindown to "never" and see if it goes away (but I really don't think it will). Just rule out one thing at a time and then maybe I can find a clue. Also, I have another few Macx, but I haven't tried to access the shares from them while access from another one is fouled up - I'll test that too.

 

FYI: I'm sharing everything via samba, ethernet is hard-wired, the Mac I'm using never goes to sleep...can't think of anything else helpful - ask me if something comes to mind.

 

I've never heard of an issue like this, so it's very tempting to say it must be a Mac issue!  But then I don't have any Mac's...  But I really have not heard of a problem, where drives and files become unavailable after a long enough delay.  That would clearly be a showstopper, if others were having it.

 

What may be helpful is to research the "error - 36" in Mac documents and knowledge bases.

I have the same setup and itunes has no problem waiting for the drives associated with my music share to spin up. This is using both a mac mini and a mbp.

 

 

I have actually had this problem on occasions also with my Mac.  I have not quite figured out what is the problem and have not had the time to look into it further.  I have come to the conclusion though that it IS a Mac problem, and one specific to finder.  It is not a problem with the server as you can reboot the mac and the shares/disks will come back just fine.  I think and am 80% positive that it is a "problem" with Finder.  Finder is nice and all but I am hoping it gets some more work done to itself in Snow Leopard.  Anyway, it seems like Finder is locally caching the directory structure and not forcing the drive to spin up on the server to get the new listing, which annoys the hell out of me.  I have had a couple instances where I have created new folders on a disk via telnet (cause the terminal window was open and I was already messing with the server) and then accessed the share via the smb volume mounted on the desktop.  I have had multiple cases were the new directory does not show up.  Usually what I can do is completely disconnect from the server by ejecting all the drives and then reconnect to them again.  This works for me 95% of the time and I do not have any problems.  For the other 5% I have to log out or restart to get the tower to come back and be usable.

 

So, yeah, this appears to me to be a Finder issue and not one with unRAID.  It might even be the implementation of smb that Mac OS X uses and some of its oddities.  Maybe if AFP was implemented this bug would go away.

I don't think this is a Finder problem: When it happens, I can't mount from the command line either. Rebooting the Mac doesn't help. Also, disk shares don't seem to be affected either, just user shares.

 

I've started a new thread since, based on the subject, I feel like I sort of hijacked this one (i.e. not related to iTunes). The new thread can be found here: http://lime-technology.com/forum/index.php?topic=3692.0 and it contains a little more information.