This sounds a little like my problem -- Although I'm not running RAID on the
partition in question (I do have RAID5 on the system). We are working with
approx 30G worth of files on a 50G ext2 file system -- we see random errors
that seem related to cache corruption ( ll_rw_blk.c complains)
I get these type messages in /var/adm/messages:
Aug 24 07:16:15 medusa kernel: attempt to access beyond end of device
Aug 24 07:16:15 medusa kernel: 08:71: rw=0, want=2083532925, limit=48845601
Aug 24 07:16:15 medusa kernel: dev 08:71 blksize=1024 blocknr=2083532924
sector=-127901448 size=1024 count=1
I'm running 2.2.11 with patch-2.2.12pre8.gz
________________________________________
Michael D. Black Principal Engineer
[EMAIL PROTECTED] 407-676-2923,x203
http://www.csi.cc Computer Science Innovations
http://www.csi.cc/~mike My home page
FAX 407-676-2355
----- Original Message -----
From: sean <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Tuesday, August 24, 1999 9:44 PM
Subject: difficulty writing reliably to working RAID setup
I set up a 106 gb RAID-0 machine using Linux software raid with the 2.2.10
kernel, an Adaptec 7XXX based controller on an Asus motherboard and 3 36gb
SCSI harddrives.
I write 5000 or so large graphics files (.png) to the RAID, which is
mounted as /home and it all goes well. However, then I run my checksum
script which checks all the .png files for integrity and I get:
raid0_map bug: hash->zone0==NULL for block 928493773
Bad md_map in ll_rw_block
repeated a few times with different block numbers.
I rerun the test and get the same errors with different block numbers,
about 15 times. Then I stop getting errors. IS the kernel simply being
overly verbose? What is the problem here? Can somebody shed some light on
t his?
Thank you!
Sean Harper
