xserve - intermittent permission problem

[paziu]

Hello everybody,

I am wondering if someone would have an idea on what possessed - a "genius" with a ghostbuster certificate might be needed...

system:
XServe 10.5.4, active directory integrated, afp, single share, 3TB of data, running for past 3 years.

problem:
every few days or a few weeks file copy via AFP is being denied on random files - a directory with 15 video files, I can OPEN/view all of them, but if I try to copy to a local machine, I only can copy 13 out of all. the other 2 return an inefficient privilege errmsg. ( all POSIX and ACL permissions are same for all files - no difference after re-setting them ) I can only copy those files to a different location on the xserve, and than all copy successfully over AFP ).
In order to read them from the original location, i have to wait, often few hours or sometimes minutes, than it copies the files fine. - the files are not being accessed/opened for write etc by anybody else. for troubleshooting purpose I killed the time machine and the antivirus - same problem.

I did not run the permission check from the disk utility, since I am under impression that it cannot do anything else to 1-2 files other than re-writing/re-setting the permissions... plus it is a disk wide/system wide check.

I would appreciate any suggestions,

Thanks,

Mike

 

[Mactasia]

I would say you haven't got anything to lose by running the Disk Utility checks if you still haven't already done so.

Can you rule out any bad sectors on the HD? It seems quite inconsistent, so this might be a possible problem?!

 

[paziu]

Thank you for your reply,

I thought about bad sectors, the logs on the SAN do not show any, and as did mention, you can OPEN the file for viewing, but you cannot COPY anywhere but locally - 3 OSX client systems getting access denied during the COPY/PASTE operation, all can OPEN the files - and the files are random, in random subfolders, all have same sets of ACLs and POSIX - then after a couple of hours a miracle happens, and the xserve allows all COPY operations...
your suggestion to go ahead and run the permission check is something I will do today - just needed somebody to confirm this to be a good idea....

Thanks again and I will let you know about the results - since the nature of this issue, I might not be able to post the results within a short time.

Mike


EDIT: since it is not the system disk, the disk utility does not give me an option to verify the disk permissions on the affected volume - the "verify disk" returned OK - ( I wonder why this is called "verify disk" since it does not check the disk, but the filesystem structure only )

I would say you haven't got anything to lose by running the Disk Utility checks if you still haven't already done so.

Can you rule out any bad sectors on the HD? It seems quite inconsistent, so this might be a possible problem?!

 

[Mactasia]

Did you set the permissions on the Share Point from within Server Admin? If so, did you 'Propagate Permissions' from the drop down in Server Admin, File Server, Permissions, Settings button (little button near the bottom that expands in to a drop down list)? That has fixed some issues for me in the past.

Good luck with the Disk Utility checks first. Let me know how you get on!

 

[paziu]

Thanks again!

Yes, the propagation of permissions was done several times during the endless troubleshooting of the problem on both the share itself or certain "random" subdirectories. in all cases, it took either few seconds or a few hours before the affected files were able to be copied..

Thanks again!

Mike

Did you set the permissions on the Share Point from within Server Admin? If so, did you 'Propagate Permissions' from the drop down in Server Admin, File Server, Permissions, Settings button (little button near the bottom that expands in to a drop down list)? That has fixed some issues for me in the past.

Good luck with the Disk Utility checks first. Let me know how you get on!

 

[paziu]

the solution to this problem, along with solution to replace the rack-mountable pizzabox xserve system is ( in my case ) to move to linux ( Gentoo ) with AFP and ACL enabled ext3/4 filesystem, and of course integration with ActiveDirectory - all tests so far showed no difference ( for the client side ) when connecting to the new system - in about 2 weeks the nice xserve will go to the storage room... on the other hand I wish I can find out what could be the permission problem - it smells like a timing or a mailformed AFP request/response... the permission denied errormessage is usually a bogus error, an error that is being displayed instead of "..an unknown error occurred.." or "error: 0x0a02c23232 - for more info consult kernel documentation"

Mike