Kernel disk# memory allocation error
since yesterday, we experience serious problems on our MCX-environment on 10.4.8 client/server.
Short story: working with our RAID-volumes creates megabytes of error messages in the system.log like: "Kernel disk0s5 memory allocation error."
As a side effect, AFP goes up to 99.9% of CPU usage (which isn't much on a Xserve G4 1GHz anyway 🙂
This happens to all users, MCX and local users.
The errors result in corrupted Volume bitmaps. We could virtually repair our drives hourly. The error is a folder thread error in the bitmap of the respective volumes.
We have tried already:
- changing the server hardware
- repairing the volumes (Disk Utility, Disk Warrior, TechToolPro)
- using MCX and local users (even creating blank local users)
Before, we experienced some permissions problems and server slow-downs. We then checked and repaired the drives with Disk Warrior and after that deleted all .DS_Store files on our RAID volumes (as a maintenance routine).
Since then, we never got rid of the problems and errors above.
Any ideas?
Xserve G4 1GHz Mac OS X (10.4.8) Xserve Raid