ktsf
09-27-2008, 06:18 AM
Hi:
I have the following MacPro (presently OS 10.5.5.) setup:
Four internal 500GB drives (Western Digital), drive A is the startup disk and contains the admin account, drive B is an emergency clone of A, C and D are stripe RAIDed for normal users. The C and D RAID is being backuped to en external firewire drive.
Every 2 weeks or so I run a smart cloning from A to B using the latest Superduper version. For normal work (i.e. no admin work) I unmount drive B to prevent it from starting up without need.
Yesterday drive A was being filled by some rsync process until it became unusable. I saw this when it was too late to do anything except killing the rsync process in Activity Monitor.
This left me with a near unusable drive A and I restarted from my emergency drive B. I searched drive A for any problematic mega-file using "WhatSize", ran repair from DiskUtility, ran DiskWarrior. Nothing changed. The drive window showed it was full up to +/- 480GB of originally 481GB. However, adding the space used by visible and invisible files gave me a number of 245GB used space, i.e. the same as before disaster struck. The space was taken up by some "dark files" that don't show up in the Finder, even when invisible files are being made visible (using TinkerTool).
In other words: I did not find a way to rescue drive A and I had to do a complete restore from drive B.
The first time this happened was in June, while I was still on OS 10.4.11
I do have the strong impression that SuperDuper might be responsible for this problem: Although I have not set it to automatic backup, it may initiate a backup procedure and, given that the target (drive B) is not mounted, it clones the source onto itself.
Log files were lost because I could not startup anymore from drive A.
Any tech hint for how to avoid this problem in the future?
I have the following MacPro (presently OS 10.5.5.) setup:
Four internal 500GB drives (Western Digital), drive A is the startup disk and contains the admin account, drive B is an emergency clone of A, C and D are stripe RAIDed for normal users. The C and D RAID is being backuped to en external firewire drive.
Every 2 weeks or so I run a smart cloning from A to B using the latest Superduper version. For normal work (i.e. no admin work) I unmount drive B to prevent it from starting up without need.
Yesterday drive A was being filled by some rsync process until it became unusable. I saw this when it was too late to do anything except killing the rsync process in Activity Monitor.
This left me with a near unusable drive A and I restarted from my emergency drive B. I searched drive A for any problematic mega-file using "WhatSize", ran repair from DiskUtility, ran DiskWarrior. Nothing changed. The drive window showed it was full up to +/- 480GB of originally 481GB. However, adding the space used by visible and invisible files gave me a number of 245GB used space, i.e. the same as before disaster struck. The space was taken up by some "dark files" that don't show up in the Finder, even when invisible files are being made visible (using TinkerTool).
In other words: I did not find a way to rescue drive A and I had to do a complete restore from drive B.
The first time this happened was in June, while I was still on OS 10.4.11
I do have the strong impression that SuperDuper might be responsible for this problem: Although I have not set it to automatic backup, it may initiate a backup procedure and, given that the target (drive B) is not mounted, it clones the source onto itself.
Log files were lost because I could not startup anymore from drive A.
Any tech hint for how to avoid this problem in the future?