From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7) Gecko/20040616 Description of problem: The sun lx50 cannot net install r4a4. It hangs at regular places in install (tcpdump trace attached) 10:46:44.680201 10.6.73.75.4227416452 > 10.6.73.218.2049: 128 read fh Unknown/1 8192 bytes @ 6578176 (DF) 10:46:44.680451 10.6.73.218 > 10.6.73.75: (frag 47310:900@7400) 10:46:44.680462 10.6.73.218 > 10.6.73.75: (frag 47310:1480@5920+) 10:46:44.680469 10.6.73.218 > 10.6.73.75: (frag 47310:1480@4440+) 10:46:44.680475 10.6.73.218 > 10.6.73.75: (frag 47310:1480@2960+) 10:46:44.680482 10.6.73.218 > 10.6.73.75: (frag 47310:1480@1480+) 10:46:44.680489 10.6.73.218.2049 > 10.6.73.75.4227416452: reply ok 1472 read (frag 47310:1480@0+) 10:46:45.380082 10.6.73.75.4227416452 > 10.6.73.218.2049: 128 read fh Unknown/1 8192 bytes @ 6578176 (DF) 10:46:45.380167 10.6.73.218 > 10.6.73.75: (frag 47311:900@7400) 10:46:45.380174 10.6.73.218 > 10.6.73.75: (frag 47311:1480@5920+) 10:46:45.380180 10.6.73.218 > 10.6.73.75: (frag 47311:1480@4440+) 10:46:45.380187 10.6.73.218 > 10.6.73.75: (frag 47311:1480@2960+) 10:46:45.380193 10.6.73.218 > 10.6.73.75: (frag 47311:1480@1480+) 10:46:45.380200 10.6.73.218.2049 > 10.6.73.75.4227416452: reply ok 1472 read (frag 47311:1480@0+) 10:46:45.975996 arp who-has 10.6.73.75 (Broadcast) tell 10.6.73.75 10:46:46.779848 10.6.73.75.4227416452 > 10.6.73.218.2049: 128 read fh Unknown/1 8192 bytes @ 6578176 (DF) ... The lx50 has incompatible bios acpi tables, so the 'acpi=off' kernel parameter must be used. Also tried using the 'noapic' parameter without effect. label lr4a4 kernel r4a4/images/pxeboot/vmlinuz append acpi=off ksdevice=eth0 console=ttyS1,9600n8 load_ramdisk=1 network ks=nfs:10.6.73.218:/storage/r4a4/ks1.cfg initrd=r4a4/images/pxeboot/initrd.img Using different nfs servers (rhas21, r4a4) resulted in hangs at slightly different places. Version-Release number of selected component (if applicable): kernel-smp-2.6.7-1.451.2.3 How reproducible: Always Steps to Reproduce: 1. pxe boot nfs install from rhas21 server (see syslinux target above) 2. nfs retries begin to occur loading rpms such as glibc 3. Actual Results: installation hangs. Expected Results: installation completes. Additional info: I'll attach a tcpdump and lspci.
Created attachment 102712 [details] lspci -v
Created attachment 102713 [details] tcpdump file
Created attachment 102714 [details] show task
This appears to be related to the bmc and ipmi setup. Disabling the ipmi lan configuration allows the installation to complete. Apparently there are specifics with adapter setup that are required for the bmc and e100 to work together. The lan setup is unexceptional: ipaddr: 10.6..73.75 netmask: 255.255.252.0 gateway: 10.6.72.1 Not finding reports on the scb2 motherboard. Earlier motherboards: http://www.uwsg.iu.edu/hypermail/linux/kernel/0110.3/1655.html
intel.com indicates that this may be fixed in e100 version 3.0.27: Subject: RE: intel scb2 server, e100 driver and ipmi interference on rhel4-alpha4 From: adaptersuppor.intel.com Version 3.0.18 was a development version. The first stable version of the 3.0.x driver was 3.0.27. Can you try that one? If the issue persists with 3.0.27, please confirm that the issue does not happen with 2.3.43.
rhel4 B2 includes the 3.0.27 e100, we will re-test there.
This bug is happening with the r3u4 beta as well. The workaround is to install with another protocol (ftp succeeded). As I recall, there have been previous problems with the bmc swallowing nfs packets, thinking they are ipmi.
Thank you for submitting this issue for consideration in Red Hat Enterprise Linux. The release for which you requested us to review is now End of Life. Please See https://access.redhat.com/support/policy/updates/errata/ If you would like Red Hat to re-consider your feature request for an active release, please re-open the request via appropriate support channels and provide additional supporting details about the importance of this issue.