My original post:
-- begin included message --
This happens in a Superworkstation machine (Sparc 2 clone), running Solaris
One of my 1 Gb Conner disks
c0t1d0 <Conner CFP1060 cyl 2329 alt 2 hd 8 sec 111>
started to make the system say:
Feb 15 10:36:35 cardeal unix: WARNING: /sbus@1,f8000000/esp@0,800000/sd@1,0 (sd1):
Feb 15 10:36:35 cardeal unix: Error for command 'read' Error Level: Retryable
Feb 15 10:36:36 cardeal unix: Requested Block 1034976, Error Block: 1034985
Feb 15 10:36:37 cardeal unix: Sense Key: Media Error
Feb 15 10:36:38 cardeal unix: Vendor 'CONNER': ASC = 0x11 (hard data error), ASCQ = 0x0, FRU = 0x15
Then to fix it I:
1) ran format/analyze/purge to repair bad blocks. It said the repairs were done.
2) made newfs - but it complained saying that there were errors on the disk.
3) so I format/repair (only repair, not format!) the bad blocks - the numbers
I got from /var/adm/messages, everytime I ran newfs.
4) repeated 2 and 3 until newfs worked.
by the end of step 4 I had a list of more than 50 bad blocks...
Then I tried to use the disk. It has only one partition. To check if everything is ok I created (ahn, tried to create) a huge file. And the system again was
complaining with the same messages above... :-(
I'm on the way to try to mark all the bad blocks as real bad blocks, but I'm
afraid that something is wrong with the disk, so I would keep marking bad blocks
Does someone ever experienced something like this? Any help will be appreciated.
Thank you very much, I'll summarize.
-- end included message --
I concluded that its really a problem in the disk. But I can manage
to use up to 80% of its full capacity, without the annoying warnings (and data
loss). I plan to continue to use it with nonsensitive data, mainly mirrors of
some ftp sites.
Many thanks to:
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:17 CDT