Script to work around incidents 1870300, 1594277 and 1753740 (DMP cannot detect re-enabled OS devices if the device names have changed on SLES 11)

  • Article ID:100032266
  • Last Published:
  • Product(s):InfoScale & Storage Foundation


This issue is specific to SLES 11 operating system. As the parameter remove_on_dev_loss of the scsi_transport_fc module has been removed in SLES 11, the OS device files are removed after a device loss with dev_loss_tmo. When the device comes back online, the device names may have changed and DMP will then not recognize its status with the
restore daemon.



Novell has released a kernel update on March 2, 2010 that allows users
to modify dev_loss_tmo value to up to 10 days.  To utilize this

1. Download and upgrade to kernel or later from

2. Create file /etc/udev/rules.d/40-rport.rules with the following
content line:

KERNEL=="rport-*", SUBSYSTEM=="fc_remote_ports", ACTION=="add",RUN+="/bin/sh -c 'echo 20 > /sys/class/fc_remote_ports/%k/fast_io_fail_tmo;echo 864000
> /sys/class/fc_remote_ports/%k/dev_loss_tmo'"

3. Reboot the system.

NOTE# Users should run:

"udevadm trigger  --action=add --subsystem-match=fc_remote_ports"

if new luns are dynamically assigned to the host.


Was this content helpful?

Get Support