Richard Troth wrote: >Two VM hosts sharing a couple of DASD strings, and a Linux v-machine >defined on both VM hosts (same minidisks), Linux will run on one >but will not run on the other. The problem is that Linux can not see >some of the minidisks. > >On the VM host where Linux comes up, the physical volumes have >two paths, both operational. On the VM host where it fails, >the physical volumes have four paths, only two are operational. >(We had a cabling issue; long story; doesn't really matter.)
We've fixed some issues in that area recently. Are you running with the latest patches applied? In particular, the fix for the problem described on developerWorks as Description: Path not operational handling. Symptom: Paths which are fine appear as not operational. Problem: When Linux encountered a deferred condition code 3, or another not operational situation, it considered the path designated in the last path used mask (lpum) to be not operational and switched it off. However, (1) the lpum shows the last *successfully* used path, and (2) deferred cc 3 doesn't necessarily mean the path is not operational. Solution: Don't switch off paths for all devices, and switch off paths for a device only according to the path not operational mask (pnom). which was dropped on 2003-02-03 for 2.4.19 and on 2003-02-20 for 2.4.7 / 2.4.17 might help with your problem. >This is not a common scenario, but one that Linux/390 should >cope with, since it should philosophically take advantage of this >fancy I/O subsystem we've got. Frankly, I don't want *my* customers >to lose a Linux volume because the operator had to vary off a couple >of paths. Varying off paths is normal stuff in big mainframe shops. Certainly. However, note that we know of certain scenarios where paths disappearing *at run time* are not always handled correctly at the moment. We're planning to fix this, but until then, please note the following restriction (from the restrictions page): * Do not re-configure or re-cable path-grouped DASDs or path-grouped channel-attached tapes while the system is running, to avoid problems with path re-validation code. In such a configuration the system may hang during re-boot. As a safe workaround use the "halt" command and perform an IPL clear. (I do not believe this applies to your situation, as the paths were not detached at run time.) Bye, Ulrich -- Dr. Ulrich Weigand [EMAIL PROTECTED]
