Technical article 104733

Getting filesystem related-panics and/or filesystem corruptions.

OpenServer 5.0.0OpenServer 5.0.2OpenServer 5.0.4OpenServer 5.0.5OpenServer 5.0.6OpenServer 5.0.7

Written · Last revised

Problem

System using SCSI hard disk(s) is panicing with filesystem related panics (for example: "HTFS: Free block <blkno> epi freed....."), or the filesystem appears to become corrupt (for example, missing files).

Cause

This may be due to read caching on the SCSI hard disks
not working correctly.  It may also be caused by improper
caching on the disk controller.

Solution

Contact the hard disk vendor to see if there is a known problem with read caching on the disk you are using, and if a firmware or microcode upgrade that corrects the problem is available.

Alternatively, you can disable read caching by editing:

/etc/conf/pack.d/Sdsk/space.c
and changing this line from:

/* If true disable target read cache on all units */ int Sdsk_set_RCD = 0;

to:

/* If true disable target read cache on all units */ int Sdsk_set_RCD = 1;

The kernel should then be relinked and the system rebooted.

Note: This is a workaround. It will reduce the performance of all the SCSI hard disks attached to the system.

These hard disks have been reported to have this problem:

Conner:       CFP-1060S

Reported on Microcode version 9WA1.60, ROM Revison 9WS020. Upgrade to Microcode version 9WA1.62 or later and SCSI ROM Revision 203C or later.

IBM:          CP30540

Reported on Revision ADB7. This is a re-badged Conner drive; corresponding Conner model number not known. Please contact IBM to confirm the problem and check if a fix is available.

Seagate:      ST3610N

Please contact Seagate to confirm the problem and check if a fix is available.

Seagate:      Barracuda 2.1GB
Reported that there is a firmware revision dated December 94
that fixes the problem. May apply to model numbers 12550N or
32550N.  Contact Seagate to confirm the problem.
HP:           D2075A

Rebadged Seagate ST3610N - please contact HP to confirm the problem and check if a fix is available.

Micropolis:   2217

Please contact Micropolis to confirm the problem and check if a fix is available.

Also, it is highly recommended that your hard disk controller,
if it does caching, be configured in write-through mode,
rather than write-back mode.  This may not be possible
in some cases, and you may have to check with the vendor
of your controller to find out if the controller supports
write-through caching.  For instance, some controllers do
not support write-through caching at certain RAID level
configurations.

From the Xinuos technical article database, record 104733. Reproduced with permission. The original is at www.sco.com/ta/104733, which is reachable but blocked from search engines. Xinuos marks this article against a retired product line. That refers to the product’s lifecycle, not to whether the advice still works — plenty of it still applies to running systems.

Next step

Need this fixed rather than explained?

Tell us the release, what the system runs and what it is doing. We will reply by email with what we would do next.

  • You deal with the people who do the work
  • No obligation, and no upgrade required before we will help
  • We tell you plainly if something is not worth doing

Send an enquiry

We reply by email. If a server is down, say so in the first line. You can also write to help@scoexpert.com.