Question

sybase iq 数据库启动报错，请高手帮助解决下。着急啊，在线等。

JN_DBW:/home/sybiq/asiq12/bin$./start_asiq @/home/sybiq/jhdw/jhdw.cfg /home/sybiq/jhdw/jhdw.db

Starting server JN_DBW_jhdw on JN_DBW at port 2640 (02/14 10:49:23)

Run Directory : /home/sybiq/asiq12/bin
Server Executable : /home/sybiq/asiq12/bin/asiqsrv12
Server Output Log : /home/sybiq/asiq12/logfiles/JN_DBW_jhdw.014.srvlog
Server Version : 12.7.0/GA
Open Client Version : 15.0/P-EBF13577 ESD #4
User Parameters : '@/home/sybiq/jhdw/jhdw.cfg' '/home/sybiq/jhdw/jhdw.db'
Default Parameters : -gn 105

=============================================================
IQ server starting with:
100 connections ( -gm )
30 cmd resources ( -iqgovern )
745 threads ( -iqmt )
512 Kb thread stack size ( -iqtss )
381440 Kb thread memory size ( -iqmt * -iqtss )
10 IQ number of cpus ( -iqnumbercpus )
=============================================================

I. 02/14 10:49:26. Adaptive Server IQ
I. 02/14 10:49:26. Version 12.7
I. 02/14 10:49:26. (64bit mode)
I. 02/14 10:49:26. Copyright 1992-2006 by Sybase, Inc. All rights reserved
I. 02/14 10:49:26.
I. 02/14 10:49:26. 10 physical processor(s) detected.
I. 02/14 10:49:26. Running on HP-UX B.11.31 U
I. 02/14 10:49:26. 49152K of memory used for caching
I. 02/14 10:49:26. Minimum cache size: 49152K, maximum cache size: 262012K
I. 02/14 10:49:26. Using a maximum page size of 4096 bytes
I. 02/14 10:49:26. Starting database "jhdw" (/home/sybiq/jhdw/jhdw.db) at Tue Feb 14 2012 10:49
I. 02/14 10:49:26. Database recovery in progress
I. 02/14 10:49:26. Last checkpoint at Mon Feb 13 2012 14:05
I. 02/14 10:49:26. Checkpoint log...
I. 02/14 10:49:33. Transaction log: /home/sybiq/jhdw/jhdw.log...

**************************************************
*** Sybase IQ Abort:
*** From: s_db.cxx:947
*** PID: 22789
*** Message: s_dbmvbio::AllocateSpecial Error
*** Thread: 856 (TID: 743)
**************************************************

I. 02/14 10:54:47. Rollback log...
I. 02/14 10:54:48. Checkpointing...
I. 02/14 10:54:48. Starting checkpoint of "jhdw" (jhdw.db) at Tue Feb 14 2012 10:54
Wrapping iqmsg: Disabled.
DBSPAWN ERROR: -80
Unable to start database server

Server failed to start (Return Code 1)

Answer 1

Sybase提供的解决方法:
强行恢复数据库步骤 （可在故障及恢复手册 第二章 75页找到）

1. 使用强制选项启动IQ服务器，在start_asiq命令中使用 -iqfrec 选项
例如：
start_asiq @query.cfg -gd dba -gm 1 -iqfrec query query.db
日志显示：
=============================================================
IQ server starting with:
200 connections ( -gm )
42 cmd resources ( -iqgovern )
1245 threads ( -iqmt )
512 Kb thread stack size ( -iqtss )
637440 Kb thread memory size ( -iqmt * -iqtss )
16 IQ number of cpus ( -iqnumbercpus )
=============================================================
I. 01/27 20:13:38. Rollback log...
I. 01/27 20:13:38. Checkpointing...
I. 01/27 20:13:38. Starting checkpoint of "query" (query.db) at Wed Jan 27 2010 20:13
I. 01/27 20:13:38. Finished checkpoint of "query" (query.db) at Wed Jan 27 2010 20:13
I. 01/27 20:13:38. Recovery complete
I. 01/27 20:13:39. Database "query" (query.db) started at Wed Jan 27 2010 20:13
I. 01/27 20:13:39. IQ Server query.
I. 01/27 20:13:39. Database server started at Wed Jan 27 2010 20:13
I. 01/27 20:13:39. Trying to start TCPIP link ...
I. 01/27 20:13:39. Starting on port 2638
I. 01/27 20:13:44. TCPIP link started successfully
I. 01/27 20:13:44. Now accepting requests
New process id is 7472
Server started successfully
你应该指定两次数据库名称，一次是指定强制恢复的数据库，一次是指定启动的数据库，-iqfrec 选项要求数据库名称。
注意这是指物理数据库的名称，是大小写敏感的。 不要使用select_dbname得到的名称，那是逻辑名称。

2. 运行sp_iqcheckdb命令去检查数据库是否存在泄露的空间
连接到正在恢复的数据库。
以分配模式运行存储过程 sp_iqcheckdb：sp_iqcheckdb 'allocation database'
如果检查显示以下信息:
======================================== ============================== =====
DBCC Allocation Mode Report
======================================== ============================== =====
** DBCC Status Errors Detected *****
** Block Count Mismatch -11857364 *****
** Blocks Leaked 11857364 *****
则说明存在泄露空间，你需要正常的停止IQ服务器，并使用-iqdroplks选项去重新启动IQ服务器 （手册77页 “在强制恢复中恢复泄露空间”）
例如：start_asiq @query.cfg -gd dba -gm 1 -iqdroplks query -iqfrec query query.db
连接到正在恢复的数据库。
以分配模式运行存储过程 sp_iqcheckdb：sp_iqcheckdb 'allocation database'
如果没有报错并且 sp_iqcheckdb 显示消息：
======================================== ============================== =====
DBCC Allocation Mode Report
======================================== ============================== =====
DBCC Status Freelist updated
DBCC Status No Errors Detected
则说明泄露的空间已恢复并且是强制执行的恢复。
发出检查点。checkpoint
3.使用常规方法停止服务器。（取消 -iqfrec 及 -iqdroplks 启动选项）
使用常规方法重新启动服务器，然后继续正常的处理。