VMware Cloud Community
69rodders
Contributor
Contributor

NetWare 6.5 SCSIHD.CDM tofactor parameter

Hi

I see on the latested SCSIHD.CDM driver available from Novell that a new parameter, tofactor, which is used as a multiplier for SCSI requests to lengthen the timeout value in the HACB. This is to multiply the time out by whatever the setting is.

I have not however been able to find any recommendations of how to use this, and whether it is still recommended to use the "retries" parameter.

Is anyone aware of the current recommendations.

Kind regards

Rodders

Tags (1)
0 Kudos
2 Replies
chucks0
Enthusiast
Enthusiast

I have been unable to find a combination of settings which avoids disk deactivation during some SAN failovers. I have tried RETRIES=6 and TOFACTOR=3, but neither seem to work. I just opened a service ticket with Novell on this issue and will post back when I get an answer.

Hi

I see on the latested SCSIHD.CDM driver available from Novell that a new parameter, tofactor, which is used as a multiplier for SCSI requests to lengthen the timeout value in the HACB. This is to multiply the time out by whatever the setting is.

I have not however been able to find any recommendations of how to use this, and whether it is still recommended to use the "retries" parameter.

Is anyone aware of the current recommendations.

Kind regards

Rodders

0 Kudos
chucks0
Enthusiast
Enthusiast

If you are running Netware 6.5 under VMware and want to adjust the disk timeout (like in Windows) so that your volumes do not deactivate during a SAN failover, here is everything you should need:

  1. Install the latest version of SCSIHD.CDM which fixes an issue which was preventing the RETRIES parameter from working (Version 3.03.10).

  2. Set the appropriate number of retries via SCSIHD RETRIES=x in the startup.ncf on the server.

    • The default disk timeout with the updated SCSIHD.CDM is approximately 73 seconds (40 seconds + 3 retries of around 11 seconds each)

    • Each additional retries will add around 10-11 seconds to the timeout.

For example, we wanted a 120 second timeout so we have SCSIHD RETRIES=8 in startup.ncf. This gives us a timeout of around 124 seconds.

0 Kudos