HDR还原主机环境问题

dellmail

回复 9# 3sane


    3q~very much!

dellmail

回复 8# liaosnet
按照您的方法，我试着进行如下实验：
假设A发生故障：   
      我试着先将A机服务器关闭、并网络断开  ；同时在B机上增加A的网络地址，并执行A的环境变量启用数据库---模拟B完全替代A工作一段时间；
假设A恢复正常后：
     将B之前增加的A地址down掉，并切换B自身的环境变量启用数据库，切换为standard状态，备份日志后 onmode -d primary A
     A启用原来网络地址并切换至informix用户 oninit -PHY 后变为fastrecovery状态，执行onmode -d secondary B 后数据库状态开始还是readonly（sec）的，可过了不到10秒，A数据库就down掉了
      无法恢复原来的A主机，B从机的状态
    何解？？？？
[informix@ontest2 ~]$ oninit -PHY
[informix@ontest2 ~]$ onstat -

IBM Informix Dynamic Server Version 11.50.FC5     -- Fast Recovery -- Up 00:00:15 -- 149192 Kbytes

[informix@ontest2 ~]$ onmode -d secondary ontest1

[informix@ontest2 ~]$ onstat -

IBM Informix Dynamic Server Version 11.50.FC5     -- Read-Only (Sec) -- Up 00:00:54 -- 157384 Kbytes

[informix@ontest2 ~]$ onstat -
shared memory not initialized for INFORMIXSERVER 'ontest2'
[informix@ontest2 ~]$ onstat -m

shared memory not initialized for INFORMIXSERVER 'ontest2'

Message Log File: /home/informix/tmp/online.log
02:02:33  Log Record: log = 86, pos = 0x1e7bc, type = OLDRSAM:ADDITEM(2, trans = 18
02:02:33  Rollforward of log record failed. iserrno = 105
02:02:33  Log Record: log = 86, pos = 0x1e7bc, type = OLDRSAM:ADDITEM(2, trans = 18
02:02:33  Logical Log 86 Complete, timestamp: 0x3f07d.
02:02:33  Assert Failed: INFORMIX-OnLine Must ABORT
        Critical media failure.
02:02:33  IBM Informix Dynamic Server Version 11.50.FC5   
02:02:33   Who: Session(16, informix@ontest2, 0, 0x4b81f470)
                Thread(34, xchg_1.1, 4b7e3708, 1)
                File: rsmirror.c Line: 1726
02:02:33  stack trace for pid 12689 written to /home/informix/tmp/af.40a9930
02:02:33   See Also: /home/informix/tmp/af.40a9930, shmem.40a9930.0
02:02:37  Starting crash time check of:
02:02:37  1. memory block headers
02:02:37  2. stacks
02:02:37  Found bad stack; mt_stack address:4c2fde08
02:02:37  rsmirror.c, line 1726, thread 34, proc id 12689, INFORMIX-OnLine Must ABORT
        Critical media failure..
02:02:41  The Master Daemon Died
02:02:41  PANIC: Attempting to bring system down

   

dellmail

A主机：
[informix@ontest1 ~]$ onstat -m

IBM Informix Dynamic Server Version 11.50.FC5     -- On-Line (Prim) -- Up 00:38:40 -- 149192 Kbytes

Message Log File: /home/informix/tmp/online.log
02:02:23  DR: Configuration parameter values of the paired servers do not match:
              Parameter: LOGFILES
              Current server's value: 18
              Paired  server's value: 14
02:02:23  DR: Secondary server needs failure recovery

02:02:24  DR: Sending log 86, size 10000 pages, 0.35 percent used
02:02:24  DR: Sending log 87 (current), size 10000 pages, 0.28 percent used
02:02:41  DR: Receive error
02:02:41  ASF Echo-Thread Server: asfcode = -25582: oserr = 4: errstr = : Network connection is broken.
System error = 4.
02:02:41  DR_ERR set to -1
02:02:41  goto bad from hdr/hdrApply.c:1791
02:02:41  DR: Failure recovery error (14)
02:02:42  DR: Turned off on primary server
02:02:42  DR: Cannot connect to secondary server
02:07:53  DR: Cannot connect to secondary server

liaosnet

回复 13# dellmail


   1， 主备机参数不匹配
   2， 似乎逻辑日志中存在一些问题。。

3sane

回复 13# dellmail


   首先主备参数必须一致。
　如果采用DRAUTO=0的模式，接管不能直接通过改环境变量。
　B机可以用脚本hdrmkpri.sh。等Ａ机恢复后，用oninit -PHY拉起，重新定义HDR关系。
　由于改环境变量各自拉起当了一次PRIM，导致数据库日志不一致，数据库罢工了，因为无法确定谁才是正确的。