Date: Sat, 21 Dec 1996 15:31:26 -0500 (EST) From: "Roger M. Levasseur" <rml@midnight.mv.com> To: freebsd-questions@freebsd.org Subject: overlapped scsi commands Message-ID: <199612212031.PAA04064@midnight.mv.com>
next in thread | raw e-mail | index | archive | help
Well, I too am getting the "overlapped commands" hiccup on my system here. I'm running on a AHA-1542C, FreeBSD 2.1.5, and sd0 is a Quantum XP32150. Until November, I was running the Quantum externally with IDE drives (master & slave), with the primary IDE drive as the boot-up device. I was running FreeBSD 1.1 during this time. I also have internally on the SCSI bus a CD/ROM drive and a QIC-525 drive. During the later part of November, I removed the IDE drives (the system is now IDE free :-) and relocated the Quantum disk into an internal drive bay with it as the boot device, and loaded up with 2.1.5. Since that time I've received several incidents like the following; Dec 15 22:33:46 midnight /kernel: sd0(aha0:0:0): timed out Dec 15 22:33:47 midnight last message repeated 2 times Dec 15 22:33:47 midnight /kernel: sd0(aha0:0:0): ABORTED COMMAND asc:4e,0 Overlapped commands attempted Dec 15 22:33:47 midnight /kernel: , retries:1 They are very infrequent, but not desired at all. There have been suggestions about proper termination. In anticipation of doing this change-over, the new internal SCSI cable was installed some weeks ago and no problems emerged. termination+power is provided by the 8Plex cd/rom drive. The Quantum has it's termination jumper removed, but is also supplying termination power to the bus. However, it has been easy to note that looking at the archive that many instances of this similar thing deals with the 'aha' device. One post (I believe it was dated August 1996) said that it was the result of a driver problem and that it was fixed in "-current" (which ever tree that was at the time and what it has become today I do not know). So, can anybody state that this is a driver problem and if there is a resolution, which version has the fix, or is this a question of cable or termination? The problem has been a sure bet chance of happening once or twice in the 25 minutes it takes to backup my system to tape (QIC-525), though it is not limited to just that activity. Note the following: Dec 4 21:48:22 midnight /kernel: sd0(aha0:0:0): timed out Dec 4 21:48:22 midnight /kernel: sd0(aha0:0:0): timed out Dec 4 21:48:22 midnight /kernel: sd0(aha0:0:0): ABORTED COMMAND asc:4e,0 Overlapped commands attempted Dec 4 21:48:22 midnight /kernel: , retries:2 Dec 4 21:48:23 midnight /kernel: sd0(aha0:0:0): timed out Dec 4 21:48:23 midnight /kernel: sd0(aha0:0:0): ABORTED COMMAND asc:4e,0 Overlapped commands attempted Dec 4 21:48:23 midnight /kernel: , FAILURE Dec 4 21:48:23 midnight /kernel: vnode_pager_input: I/O read error Dec 4 21:48:23 midnight /kernel: vm_fault: pager input (probably hardware) error, PID 204 failure Dec 4 21:48:23 midnight /kernel: pid 204 (netscape), uid 1231: exited on signal 11 Resuming use of a netscape executable caused page-in activity that in the face of this overlapped command hiccup (this is the only time that the FAILURE was generated in all of the incidents), resulted in the process being terminated. thanks in advance, -roger -- Roger M. Levasseur rml@midnight.mv.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199612212031.PAA04064>