Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 27 Jul 2001 17:40:15 +0200 (CEST)
From:      wilko@FreeBSD.org
To:        FreeBSD-gnats-submit@FreeBSD.org
Subject:   kern/29264: Recovery from LIPs on FCAL using isp not bulletproof
Message-ID:  <200107271540.f6RFeFa13011@freebie.xs4all.nl>

next in thread | raw e-mail | index | archive | help

>Number:         29264
>Category:       kern
>Synopsis:       Recovery from LIPs on FCAL using isp not bulletproof
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    freebsd-bugs
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Fri Jul 27 08:50:01 PDT 2001
>Closed-Date:
>Last-Modified:
>Originator:     Wilko Bulte
>Release:        FreeBSD 4-STABLE i386
>Organization:
Private FreeBSD site - The Netherlands
>Environment:

        FreeBSD-stable as of jul 27, 2001. (Old & weary) PTI ISP-2100-based
	FCAL adapter.

>Description:

        Forcing LIPs on the FCAL will after some time result in the 
	FCAL disk being dropped (invalidating pack). This will take
	quite some clobbering of the AL though. And is supposedly
	also dependent on the adapter / firmware in use.

Jul 27 14:55:34 p2 /kernel: isp0: Name Server Database Changed
Jul 27 14:55:34 p2 /kernel: isp0: Firmware State <Config Wait->Ready>
Jul 27 14:55:34 p2 /kernel: isp0: Loop ID 109, AL_PA 0x29, Port ID 0x29,
Loop St
ate 0x2, Topology 'Private Loop'
Jul 27 14:55:34 p2 /kernel: isp0: Target 109 (Loop 0x6d) Port ID 0x29 (role
Init
iator) Arrived
Jul 27 14:55:34 p2 /kernel: Port WWN 0x210000e08b000174
Jul 27 14:55:34 p2 /kernel: Node WWN 0x200000e08b000174
Jul 27 14:55:34 p2 /kernel: isp0: Retaining Loop ID 0x0 for Target 0 (Port
0xef)
Jul 27 14:55:34 p2 /kernel: isp0: Retaining Loop ID 0x7d for Target 125
(Port 0x
1)
Jul 27 14:56:34 p2 /kernel: isp0: command timed out for 0.125.0
Jul 27 14:56:35 p2 last message repeated 15 times
Jul 27 14:56:38 p2 /kernel: (da1:isp0:0:125:0): isp0: watchdog timeout (2b,
0)
Jul 27 14:56:38 p2 /kernel: isp0: cannot find handle 0x2b in xflist
Jul 27 14:56:38 p2 /kernel: (da1:isp0:0:125:0): isp0: watchdog timeout (2c,
0)
Jul 27 14:56:38 p2 /kernel: isp0: cannot find handle 0x2c in xflist
Jul 27 14:56:38 p2 /kernel: (da1:isp0:0:125:0): isp0: watchdog timeout (2d,
0)
Jul 27 14:56:38 p2 /kernel: isp0: cannot find handle 0x2d in xflist
Jul 27 14:56:38 p2 /kernel: (da1:isp0:0:125:0): isp0: watchdog timeout (2e,
0)
Jul 27 14:56:38 p2 /kernel: isp0: cannot find handle 0x2e in xflist


>How-To-Repeat:

        Inject LIPs during heavy disk I/O every 180 seconds did this.
	My heavy I/O was a buildworld with /usr/src and /usr/obj each
	on their own FC disk, both on the same loop.

>Fix:

	mjacob has sysctl changes sitting in the pipeline to allow 
	retry tuning for the da(4) driver. Adding delay hysteresis
	(Matt's words) in the isp driver is also called for.
>Release-Note:
>Audit-Trail:
>Unformatted:

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200107271540.f6RFeFa13011>