Consistent Kernel Panic with OVS 2.2.1 / OEL 4U8 PVM template / perl
zsysdbaFeb 17 2011 — edited Feb 23 2011I am getting a consistent kenel panic running OEL 4U8 PVM as a 10.2.0.3 db host. From another post, I was able to configure the VM to get crash dumps before restart. I'm using the strings command on the core dump to search for the panic output. Here is an example.
<1>Unable to handle kernel paging request at ffffff807a936000 RIP:
<1><ffffffff801ebd22>{clear_page+18}
<4>PML4 103067 PGD c5b067 PMD e30067 PTE 800000007a936167
<0>Oops: 0003 [1] SMP
<4>CPU 1
<4>Modules linked in: md5 ipv6 oracleasm(U) autofs4 sunrpc dm_mirror dm_mod xennet xenblk ext3 jbd
<4>Pid: 27817, comm: perl Not tainted 2.6.9-89.0.0.0.1.ELxenU
<4>RIP: e030:[<ffffffff801ebd22>] <ffffffff801ebd22>{clear_page+18}
<4>RSP: e02b:ffffff807542dca0 EFLAGS: 00010216
<4>RAX: 0000000000000000 RBX: ffffff807a936000 RCX: 000000000000003f
<4>RDX: 000000007a936000 RSI: 0000000000000000 RDI: ffffff807a936000
<4>RBP: ffffff8048ae2dc0 R08: 0000000000000000 R09: ffffff8048ae2dc8
<4>R10: 0000000000000000 R11: 0000000000000000 R12: ffffff8048ae2dc0
<4>R13: 00000000ffffc000 R14: ffffff800929e018 R15: 00000000ffffc000
<4>FS: 0000002a970596e0(0000) GS:ffffffff80439f00(0000) knlGS:0000000000000000
<4>CS: e033 DS: 002b ES: 002b
<4>Process perl (pid: 27817, threadinfo ffffff807542c000, task ffffff8012ef47f0)
<4>Stack: ffffffff8015965b ffffff8048ae2e38 ffffffff80163929 ffffff800929e018
<4> ffffff8048ae2dc0 00000000ffffc000 ffffff8014316140 ffffff80018b7f98
<4> ffffffff8018104c 000000000000001e
<4>Call Trace:<ffffffff8015965b>{get_zeroed_page+104} <ffffffff80163929>{__pmd_alloc+39}
<4> <ffffffff8018104c>{install_arg_page+167} <ffffffff80127b49>{ia32_setup_arg_pages+466}
<4> <ffffffff80128380>{load_elf32_binary+2065} <ffffffff801a2c86>{load_elf_binary+5452}
<4> <ffffffff80157d02>{generic_file_aio_read+48} <ffffffff80177711>{do_sync_read+178}
<4> <ffffffff80127b6f>{load_elf32_binary+0} <ffffffff80182892>{search_binary_handler+209}
<4> <ffffffff8019f4e5>{compat_do_execve+398} <ffffffff80120007>{sys32_execve+53}
<4> <ffffffff8011e27d>{ia32_ptregs_common+37} <ffffffff8011e074>{ia32_syscall+0}
<4>
<4>Code: 48 89 07 48 89 47 08 48 89 47 10 48 89 47 18 48 89 47 20 48
<1>RIP <ffffffff801ebd22>{clear_page+18} RSP <ffffff807542dca0>
<0>CR2: ffffff807a936000
<4> <0>Kernel panic - not syncing: Oops
I've captured about 5 of these so far and the panic is always "Unable to handle kernel paging request" and the process is always perl. I suspect it is the 10.2.0.3.0 Oracle agent as this runs as perl
ps -ef|grep perl
oracle 7287 1 0 Feb16 ? 00:00:07 /opt/oracle/product/10.2.0/grid1/agent10g/perl/bin/perl /opt/oracle/product/10.2.0/grid1/agent10g/bin/emwd.pl agent
Can anyone point me to how I would use the crash command in dom-0 to debug this further? Where do I get the required "mapfile" and "namelist" files the crash command wants?