SUMMARY: Recovery from interrupted SAN connection

From: Koef <>
Date: Sat Oct 07 2006 - 12:22:04 EDT
Original question below. It was suggested not to put essential filesystems
on SAN connected storage. There was a hint that changes to some <driver>.conf
could help a filesystem to recover after a loss of SAN connectivity, but no
details were given. Thanks to all responders.


On Wed, Oct 04, 2006 at 08:29:22PM +0200, Koef wrote:
> I have a Solaris 8 machine connected via a single Qlogic fibre HBA to EMC
> storage. When the SAN connection is lost, i.e. when pulling the fibre
> connector, following is logged at the console:
> WARNING: md: d2: write error on 
>          /dev/dsk/c0t6006048000028775062953594D433430d0s4
> WARNING: /scsi_vhci/ssd@g6006048000028775062953594d433431 (ssd1):
>          ssdrestart transport failed (fffffffe)
> WARNING: /scsi_vhci/ss
>          d@g6006048000028775062953594d433432 (ssd0):
>          transport rejected (-2)
> WARNING: ufs log for /var changed state to Error
> WARNING: Please umount (1M) /var and run fsck(1M)
> When I plug the fibre connector back in, the /var filesystem, which is on
> the SAN, does not recover. I cannot even login on the console:
> console login: root
> Password:
> No utmpx entry. You must exec "login" from the lowest level "shell".
> INIT: failed write of utmpx entry:"co"
> INIT: failed write of utmpx entry:"co"
> console login:
> To regain access to the machine at this point it must be sent a "break"
> signal and then booted from the "ok" prompt. It then drops single user
> complaining about bad super blocks in the SAN mounted filesystems.
> Question: is there a more elegant way to recover from, or to prepare for,
> single SAN fibre link interruptions? Thank you.
> -- 
> Koef.
> _______________________________________________
> sunmanagers mailing list
sunmanagers mailing list
Received on Sat Oct 7 12:22:47 2006

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:44:01 EST