[OpenAFS-devel] CopyOnWrite failure and EINTR

[Ken Aaker]

Hello,

I have a situation where one volume seems to be repeatedly destroyed. 
Nothing major, just the volume that my home directory is on ;-) After 
digging around for a while, I think I found a message that gives me a 
lead about what is going on.

In /usr/afs/logs/Filelog I found a message about a CopyOnWrite failure. 
If I'm reading the message properly, the FDH_READ() call is getting in 
EINTR rc. I'm puzzled about this, because according to one book, the 
read call that is inside the FDH_READ() macro should be restarted, if an 
EINTR occurs, and according to a couple of articles on the web, the read 
should be checked for EINTR and restarted where it stopped. I don't have 
any clues yet about why the read is being interrupted..... The volume 
that's always damaged is home.kdaaker, which then seems to destroy 
home.kdaaker.backup when the bos salvage is done??? (Damage being 
everything in the first level directory on the volume becomes an orphan).

File Server System configuration is

Dell 1400 Server with afs partitions on a MEGARAID controller (2 20GB 
partitions /vicepa /vicepb).
3 36MB SCSI Drives in a RAID 5 configuration.
800MHz Pentium III

Linux 2.4.2 SMP kernel (Caldera Distribution).

openafs 1.2.2 built on the system.

------------The volume list ---------------------------------------
Total number of volumes on server lars partition /vicepa: 12
common.downloads                  536870942 RW    3312560 K On-line
home.ccain                        536870936 RW      81229 K On-line
home.ccain.backup                 536870938 BK      81229 K On-line
home.kjduncan                     536870933 RW     673740 K On-line
home.kjduncan.backup              536870935 BK     673740 K On-line
i386_linux24                      536870918 RW     217840 K On-line
i386_linux24.usr                  536870921 RW          3 K On-line
i386_linux24.usr.afsws            536870924 RW      13510 K On-line
root.afs                          536870912 RW          4 K On-line
root.afs.readonly                 536870913 RO          4 K On-line
root.cell                         536870915 RW          5 K On-line
root.cell.readonly                536870916 RO          5 K On-line

Total volumes onLine 12 ; Total volumes offLine 0 ; Total busy 0

Total number of volumes on server lars partition /vicepb: 10
common                            536870939 RW     243659 K On-line
common.readonly                   536870940 RO     243659 K On-line
home                              536870927 RW          7 K On-line
home.backup                       536870929 BK          7 K On-line
home.jengel                       536870945 RW     112046 K On-line
home.jengel.backup                536870947 BK     112046 K On-line
home.kdaaker                      536870951 RW     587584 K On-line
home.kdaaker.backup               536870953 BK     587672 K On-line
home.testuser                     536870948 RW          7 K On-line
home.testuser.backup              536870950 BK          7 K On-line

Total volumes onLine 10 ; Total volumes offLine 0 ; Total busy 0

----------------------------------------------------------


-- 
work -> kenaaker@silverbacksystems.com	(507) 289-6910 ext 1
home -> kenaaker@qwest.net