Description of problem: After installation of the RHEL 5.4 RC2 O/S, setting up multipathing, the server reboots successfully. Upon adding additional mass storage arrays to the DOM0, the server does not survive a reboot. It stops at the ELILO bootloader and does not load the RHEL kernel. NOTE: Issue occurs on both Emulex and QLogic HBAs. NOTE: Same issue was also observed after installing xen-ia64-guest-firmware-1.0.13-1-ia64.rpm with external storage attached with emulex and qlogic hbas. Version-Release number of selected component (if applicable): RHEL 5.4 RC 2 How reproducible: This problem is consistently reproducible. Steps to Reproduce: 1. Present a boot LUN from an external storage array. 2. Install the RHEL O/S with virtulization, set up multipathing and reboot to observe that the system comes back up. 3. Present LUNs from additional external storage arrays, with a few LUNs each and verify the multipathing. 4. Reboot the server to verify that it does not get past the ELILO prompt. Actual results: The ELILO boot loader loads, the initrd completes, vmlinux kernel completes then the system hangs and does go any further. Expected results: After adding additional external storage to the system, rebooting, I would expect that the system would boot the O/S and present the login screen. Additional info: Server: BL860c Array: EVA-XL 8000 (boot LUN)
Do you know if this used to work? That is, is this a regression from 5.3? Also, I don't understand this statement: NOTE: Same issue was also observed after installing xen-ia64-guest-firmware-1.0.13-1-ia64.rpm with external storage attached with emulex and qlogic hbas. What does booting the dom0 have to do with the xen-ia64-guest-firmware? Or are you talking about booting a fully virtualized guest? Chris Lalancette
>>Do you know if this used to work? That is, is this a regression from 5.3? We have not tested 5.3 or 5.2 with IA64. >>NOTE: Same issue was also observed after installing >>xen-ia64-guest-firmware-1.0.13-1-ia64.rpm with external storage attached with >>emulex and qlogic hbas What we did was reinstall rhel 5.4 IA64. We noticed this rpm, xen-ia64-guest-firmware-1.0.13-1-ia64.rpm, and thought that it might solve our problem. After installing this rpm, we rebooted the dom0, and experienced the same behavior as was seen with adding the external storage. >>What does booting the dom0 have to do with the xen-ia64-guest-firmware? Or >>are you talking about booting a fully virtualized guest? We thought this might have fixed our problem as described in the question above. This has nothing to do with guests. We have no log files to attach as we can not boot the dom0. After experiencing this issue, we tried to boot the base kernel and not the xen kernel, and still could not boot.
(In reply to comment #2) > >>Do you know if this used to work? That is, is this a regression from 5.3? > We have not tested 5.3 or 5.2 with IA64. OK. If you have access to it, it would be very interesting to try out 5.3 and see if it works there. > > >>NOTE: Same issue was also observed after installing > >>xen-ia64-guest-firmware-1.0.13-1-ia64.rpm with external storage attached with > >>emulex and qlogic hbas > What we did was reinstall rhel 5.4 IA64. We noticed this rpm, > xen-ia64-guest-firmware-1.0.13-1-ia64.rpm, and thought that it might solve our > problem. After installing this rpm, we rebooted the dom0, and experienced the > same behavior as was seen with adding the external storage. > > >>What does booting the dom0 have to do with the xen-ia64-guest-firmware? Or >>are you talking about booting a fully virtualized guest? > We thought this might have fixed our problem as described in the question > above. > This has nothing to do with guests. OK, that makes a lot more sense. The xen-ia64-guest-firmware package is used strictly for booting Xen IA64 fully virtualized guests (and it's required to do that). It has nothing to do with dom0 booting, as you found out. > > We have no log files to attach as we can not boot the dom0. Can you hook up a serial console and get at least basic booting information out of the hypervisor and the kernel? It's kind of hard to debug flying blind. > > After experiencing this issue, we tried to boot the base kernel and not the xen > kernel, and still could not boot. OK, that's interesting information. I'm going to flip it over to kernel then, as this might be some kind of problem (or known issue) with the storage. If it turns out to be Xen specific, please feel free to flip it back to kernel-xen. Chris Lalancette
Closing this old XEN bug.