lauren@vortex.COM (Lauren Weinstein) (11/27/90)
Greetings. Has anyone else seen this message from the system under
ix2.0.2?
------
*** DEVICE ERROR: Medium has been changed in drive ***
*** Controller 1 (Adaptec AHA-1540 SCSI Host Adapter), DISK Drive 0,
Absolute Sector # 43014 ***
------
This was for a 660MB disk on the 1540. The disk had been running
fine up to this point (as long as certain I/O cards which the 1540
doesn't like weren't in the system, but that's a story for another
time) then this message popped up out of the blue last night.
My first thought was that a sector remapping had occurred, but the
grown list for the disk shows no changes.
Does this look familiar to anyone? Thanks.
--Lauren--minow@mountn.dec.com (Martin Minow) (11/28/90)
In article <66.UUL1.3#1@vortex.COM> lauren@vortex.COM (Lauren Weinstein) writes: >*** DEVICE ERROR: Medium has been changed in drive *** ... >Does this look familiar to anyone? Thanks. Hi Lauren, long time no see. There is a "medium changed" sense code that the drive can return. It really only is relevant to floppy and cd-rom devices that are under operator control. It is possible that your drive/controller returns that when the processor executes a mode select (that changes drive parameters), but that doesn't accord with my understanding of SCSI-II. If you can get a dump of the actual extended sense data block (from the error log?), I could take a swing at interpreting it (the interpretation should be in your drive's manual, or in the SCSI-II spec if your drive supports the common command set). For now, I'd treat it as a non-serious glitch. Martin Minow minow@bolt.enet.dec.com
garyb@SSD.CSD.HARRIS.COM (Gary Barton) (11/28/90)
In article <66.UUL1.3#1@vortex.COM> lauren@vortex.COM (Lauren Weinstein) writes: >*** DEVICE ERROR: Medium has been changed in drive *** Most CCS disk drives use the same additional sense code to indicate a device was reset or the media may have changed. I suspect that something caused the SCSI bus to be reset, which then generates a check condition with this error code on the next access. SCSI device drivers generally expect to see this condition after they have forced a device or bus reset, but if it occurs outside of their control, they treat it like any other error. It could be that device/bus reset was attempted as part of some error recovery procedure by the disk (or other) driver. Device themselves will sometimes cause a bus reset. I have seen some disk drives (Fujitsu) with strappings that tell the imbedded SCSI controller whether or not it should activate the SCSI bus reset when the drive is powered on. The SCSI reset line could in theory be activated by plugging/unplugging a device from an operating SCSI bus, or perhaps there could be some other cabling/termination problem. The point is, this is most likely the result of device or bus reset, not a media changed condition. -- +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ | Gary Barton | System Software Development | | Lead Engineer | Harris Computer Systems Division | | garyb@hcx1.csd.ssd.harris.com | 2101 W. Cypress Creek Rd. | | gbarton@ssd.harris.com | Ft. Lauderdale. FL 33309 | | uunet!hcx1!garyb | (305) 974-1700 | +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
jgd@rsiatl.UUCP (John G. DeArmond) (11/28/90)
garyb@SSD.CSD.HARRIS.COM (Gary Barton) writes: >In article <66.UUL1.3#1@vortex.COM> lauren@vortex.COM (Lauren Weinstein) writes: >>*** DEVICE ERROR: Medium has been changed in drive *** >Most CCS disk drives use the same additional sense code to indicate a >device was reset or the media may have changed. I suspect that >something caused the SCSI bus to be reset, which then generates a >check condition with this error code on the next access. SCSI device >drivers generally expect to see this condition after they have forced >a device or bus reset, but if it occurs outside of their control, they >treat it like any other error. It could be that device/bus reset was >attempted as part of some error recovery procedure by the disk (or >other) driver. I think this is pretty much correct in the case of ISC's handler. I've seen that message here in the context of something really gross happening to the SCSI bus - cable working loose or a radio being keyed too close to the cable or a lightening strike nearby or a drive giving it the last dying quiver :-( Unless it repeats, I'd write it off to cosmic causes. When I had a drive dying, I saw that a lot. About the only recourse is to pull the plug and let Mr. fsck cleanup the carnage. The ISC driver just says "to hell with it, I'm leaving" and that's that. John -- John De Armond, WD4OQC | "Purveyors of speed to the Trade" (tm) Rapid Deployment System, Inc. | Home of the Nidgets (tm) Marietta, Ga | {emory,uunet}!rsiatl!jgd | "Vote early, Vote often"
loc@tmsoft.UUCP (Leigh Clayton) (11/29/90)
<lauren@vortex.COM (Lauren Weinstein) > asks about
------
*** DEVICE ERROR: Medium has been changed in drive ***
*** Controller 1 (Adaptec AHA-1540 SCSI Host Adapter), DISK Drive 0,
Absolute Sector # 43014 ***
------
I get this message frequently, as a result of spinning down/up (and maybe
changing) the Magneto-optical disk I use as Drive 1 (also 386/ix 2.0.2).
It is apparantly caused by the drive telling the SCSI adapter that I've
changed the platter, in the hopes that the adapter will tell the driver and the
driver will refresh it's notions about what is spinning there. Of course
386/ix does *not* do this ... it just prints the message.
I have no idea why your disk should be sending this SCSI signal, though. If
I had to guess, I'd guess that something like a power or logic transient was
causing your disk (or maybe the Adaptec card) to power down/up occasionally.
-----------------------------------------------------------
loc@tmsoft.UUCP (Leigh Clayton) uunet!mnetor!tmsoft!locwcarroll@encore.com (Mr. New Dad) (11/29/90)
From article <1604@travis.csd.harris.com>, by garyb@SSD.CSD.HARRIS.COM (Gary Barton): > In article <66.UUL1.3#1@vortex.COM> lauren@vortex.COM (Lauren Weinstein) writes: > >>*** DEVICE ERROR: Medium has been changed in drive *** > > Most CCS disk drives use the same additional sense code to indicate a > device was reset or the media may have changed. CCS specifies sense code 0x28 as "Media has changed" and sense code 0x29 as "Device has been reset". If you have a drive that uses ASC 0x28 to indicate a reset, then it does not follow the CCS spec and should not be advertised as a CCS drive. I have never encountered a drive that displays the behavior you state. -- William R. Carroll (Encore Computer Corp., Ft. Lauderdale FL) wcarroll@encore.com uunet!gould!wcarroll "The brain-dead should not be allowed to operate motor vehicles!" - Me
src@scuzzy.in-berlin.de (Heiko Blume) (12/02/90)
wcarroll@encore.com (Mr. New Dad) writes: >CCS specifies sense code 0x28 as "Media has changed" and sense code >0x29 as "Device has been reset". If you have a drive that uses ASC 0x28 >to indicate a reset, then it does not follow the CCS spec and should not >be advertised as a CCS drive. >I have never encountered a drive that displays the behavior you state. i once connected a seagate st-296n disk along with two CDC wren-III disks to a adaptec aha-1540a run by isc 2.2 and got the media changed error. however that same seagate disk works with a wren-V and a Quantum pro80s on my machine under isc 2.2. looks like the cheapo seagate caused the problem. a newer seagate now works ok with the two wren-IIIs, probably a newer revision from seagate. -- Heiko Blume <-+-> src@scuzzy.in-berlin.de <-+-> (+49 30) 691 88 93 public source archive [HST V.42bis]: scuzzy Any ACU,f 38400 6919520 gin:--gin: nuucp sword: nuucp uucp scuzzy!/src/README /your/home
garyb@SSD.CSD.HARRIS.COM (Gary Barton) (12/06/90)
In article <129885@jake.encore.com> wcarroll@encore.com (Mr. New Dad) writes: >> >> Most CCS disk drives use the same additional sense code to indicate a >> device was reset or the media may have changed. >> >CCS specifies sense code 0x28 as "Media has changed" and sense code >0x29 as "Device has been reset". If you have a drive that uses ASC 0x28 >to indicate a reset, then it does not follow the CCS spec and should not >be advertised as a CCS drive. > I stand corrected with regard to my statement that most CCS drives exhibit this behavior. I should have said that some drives whose vendors claim CCS or SCSI-2 compliance actually use the ASC 0x29 to indicate both conditions, with ASC 0x28 reserved, or assigned to a different error condition. Both CCS recommendations and the preliminary SCSI-2 standard indicate that ASC 0x28 is assigned to "medium changed", and ASC 0x29 is assigned to a "reset" condition. >I have never encountered a drive that displays the behavior you state. > The Cipher st150s series of 1/4" cartridge tape drives cannot distinguish between these two sense conditions. The Cipher model st150s2 (so named for it's supposed SCSI-2 compliance) uses ASC 0x29 for both conditions, with ASC assigned to an entirely different error. However, an examination of a couple of other removable media devices (Maxtor Tahiti, Archive cartridge tape, etc) yields behavior that does in fact conform to the standards. I do recall seeing other devices that exhibit this objectionable behavior, but I can't remember exactly which units they were. I guess I have seen too many OEM manuals over time, and now they are all running together. I'm not sure if Cipher's embedded controller was their own design, but if it was not, you can bet that it can be found in other cartridge tape drives as well, and those units will most likely exhibit this same behavior. +--------------------------------------------------------------+ | Gary Barton | Systems Software Development | | Lead Engineer | Harris Computer Systems Division | | garyb@csd.ssd.harris.com | Ft Lauderdale, Florida | | uunet!hcx1!garyb | (305) 974-1700 | +--------------------------------------------------------------+