有关aio引起AI宕机的coredump分析.docxVIP

有关 aio 引起 AIX 宕机的 core_dump 分析2008.03.27 前些日子,客户的 S7A 主机发生了几次宕机,产生了 CORE_DUMP 文件,下面是利用 crash 命令分析宕机原因的过程 pwd / # hostname s7a01 # cd /var/adm/ras # ls -l 查看 core 文件名称total 395133 -rw-rw-r-- 1 root system 4226 Apr 02 2003 BosMenus.log -rw-r--r-- 1 root system 2 Jan 07 2000 SRCSemID -rw 1 root system 8192 May 20 13:35 bootlog -rw-r--r-- 1 root system 8388 Apr 02 2003 bosinst.data -rw-rw-r-- 1 root system 16384 Apr 02 2003 bosinstlog --w 1 root system 2 May 16 15:47 bounds -rw-r--r-- 1 bin bin 197206 Jan 01 1970 codepoint.cat -rw--w--w- 1 root system 16384 May 20 15:52 conslog --w 1 root system 21 May 16 15:47 copyfilename -rw-r--r-- 1 root system 57078 Apr 02 2003 devinst.log -rw-r--r-- 1 root system 83319 May 20 14:00 diag_log -rw 1 root system 8192 May 16 15:49 dumpsymplog -rw-r--r-- 1 root system 151552 May 20 15:52 errlog -rw-r--r-- 1 root system 151552 Apr 22 2004 errlog0422.log -r--r--r-- 1 bin bin 103968 Jan 07 2000 errtmplt -rw-r--r-- 1 root system 7949 Apr 02 2003 image.data -rw-r--r-- 1 root system 8192 May 20 13:21 nimlog -rw-rw-rw- 1 root system 1334264 Jan 20 2000 trcfile -rw 1 root system 200136704 May 16 15:47 vmcore.0 # crash vmcore.0 开打 vmcore.0 文件 Using /unix as the default namelist file. 2 dump routines failed. The following were recorded: 0x0141cbe8 failed with rc=14 0 failed with rc=14 stat 查看宕机时的状态sysname: AIX nodename: s7a01 release: 3 version: 4 machine: 000AAD014C00 time of crash: Tue May 16 15:05:18 TAIST 2006 age of system: 22 hr., 51 min. xmalloc debug: disabled abend code: 300 查看错误代码,这个代码很关键 csa: 0x2ff3b400 exception struct: dar: 0dsisr: 0 srv: 0dar2: 0dsirr: 0 (errno) Error 0 trace -m Skipping first MST MST STACK TRACE: 0x2ff3b400 (excpt0a00000000000004 (intpri=11) IAR: .compare_and_swap+2c (0000a4ec): stw r9,0x0(r4) LR: .[aiopin:untie_knot]+a8 (0143d7a8) 2ff3a2e0: .[aio.ext:qlioreq]+b0 (014376ec) 2ff3a340: .[aio.ext:listio]+128 (01438f5c) 2ff3