oracle启动卡住了,hsdb启动卡住问题

AS数据库连接不上问题，中间件的启动界面停止在如下位置：[30706]Plugin[com.hundsun.fbase.security] init ...done!

[30706]Plugin[com.hundsun.fbase.hsdb] init ...

然后gdb进程查看线程堆栈，如下所示：

(gdb) t 1

[Switching to thread 1 (Thread 4160174688 (LWP 13126))]#0  0xffffe410 in

__kernel_vsyscall ()

(gdb) where

#0  0xffffe410 in __kernel_vsyscall ()

#1  0x0070b3a2 in times () from /lib/tls/libc.so.6

#2  0xf762a651 in sltrgatime64 () from

/u01/app/oracle/product/10.2.0/db_1/lib32/libclntsh.so.10.1

#3  0xf712408b in kghinp () from

/u01/app/oracle/product/10.2.0/db_1/lib32/libclntsh.so.10.1

#4  0xf6d31467 in kpuinit0 () from

/u01/app/oracle/product/10.2.0/db_1/lib32/libclntsh.so.10.1

#5  0xf6d308ae in kpuinit () from

/u01/app/oracle/product/10.2.0/db_1/lib32/libclntsh.so.10.1

#6  0xf6dfc57e in OCIEnvInit () from

/u01/app/oracle/product/10.2.0/db_1/lib32/libclntsh.so.10.1

#7  0xf6cbe476 in sqlcxa () from /u01/app/oracle/product/10.2.0/db_1/lib32/libclntsh.so.10.1

#8  0xf6cbf268 in sqlcfx () from

/u01/app/oracle/product/10.2.0/db_1/lib32/libclntsh.so.10.1

#9  0xf6ca025d in sqlcmex () from

/u01/app/oracle/product/10.2.0/db_1/lib32/libclntsh.so.10.1

#10 0xf6ca0b06 in sqlcxt () from /u01/app/oracle/product/10.2.0/db_1/lib32/libclntsh.so.10.1

#11 0xf6899709 in HSQLInitHandle () from

/home/hundsun/linux.i386/lib/libfhsdb_oracle10.so

#12 0xf69162f5 in CConnectionImpl::connectionDB () from

/home/hundsun/linux.i386/lib/libfsc_f2hsdb.so

#13 0xf69148c4 in CDataSourceImpl::m_CreateConnection () from

/home/hundsun/linux.i386/lib/libfsc_f2hsdb.so

#14 0xf69146a6 in CDataSourceImpl::CreatConnectionByCount () from

/home/hundsun/linux.i386/lib/libfsc_f2hsdb.so

#15 0xf6913809 in CDataSourceImpl::CDataSourceImpl () from

/home/hundsun/linux.i386/lib/libfsc_f2hsdb.so

#16 0xf690e828 in CHSDBImpl::mf_CreateDataSource () from

/home/hundsun/linux.i386/lib/libfsc_f2hsdb.so

#17 0xf690e601 in CHSDBImpl::mf_CreateAllDataSource () from

/home/hundsun/linux.i386/lib/libfsc_f2hsdb.so

#18 0xf690e289 in CHSDBImpl::OnInit () from

/home/hundsun/linux.i386/lib/libfsc_f2hsdb.so

#19 0xf6910628 in HSDBSvrInit () from

/home/hundsun/linux.i386/lib/libfsc_f2hsdb.so

#20 0xf7f9d4ea in CF2CoreImpl::mf_InitPlugin () from /home/hundsun/linux.i386/lib/libfsc_f2core.so

#21 0xf7f9c08c in CF2CoreImpl::Load () from

/home/hundsun/linux.i386/lib/libfsc_f2core.so

#22 0x0804ae34 in CShell::init ()

#23 0x0804dbd8 in main ()问题原因：根据函数sltrgatime64 ()上网搜索，发现了原来是oracle客户端(版本为10.2.0.1.0)一个bug

bug描述：事实上只要Linux x86主机运行天数是是24.8的倍数都有可能引发该bug，因为time()函数值为null，造成无限死循环，从而耗尽cpu。参考文档：Doc ID: 338461.1 SQL*Plus 10.2.0.1 Hangs, When System Uptime Is Long Period of

Time

Doc ID: 4612267.8 Bug 4612267 - OCI client spins when machine uptime >= 249

days详细信息可以参考sqlplus还是可以连上数据库，根据bug描述，也应该连不上数据库才对。通过ldd查看sqlplus的依赖库，我们发现原来调用的底层库不是一个。[oracle@txas1 ~]$ ldd /u01/app/oracle/product/10.2.0/db_1/bin/sqlplus

libsqlplus.so => /u01/app/oracle/product/10.2.0/db_1/lib/libsqlplus.so

(0x0000002a95557000)

libclntsh.so.10.1 =>

/u01/app/oracle/product/10.2.0/db_1/lib/libclntsh.so.10.1 (0x0000002a95748000)

libnnz10.so

=> /u01/app/oracle/product/10.2.0/db_1/lib/libnnz10.so (0x0000002a96a96000)

libdl.so.2

=> /lib64/libdl.so.2 (0x0000003fa9b00000)

libm.so.6

=> /lib64/tls/libm.so.6 (0x0000003fa9900000)

libpthread.so.0

=> /lib64/tls/libpthread.so.0 (0x0000003fa9d00000)

libnsl.so.1

=> /lib64/libnsl.so.1 (0x0000003faf300000)

libc.so.6

=> /lib64/tls/libc.so.6 (0x0000003fa9600000)

/lib64/ld-linux-x86-64.so.2 (0x0000003fa9200000)前面堆栈显示的我们调用的库是：/u01/app/oracle/product/10.2.0/db_1/lib32/libclntsh.so.10.1

sqlplus是64位的程序，我们的程序是32位的，导致我们的问题出现。解决办法：经常重启主机，避免天数到(治标不治本)更新orcle的补丁，对该bug单独打临时patch 4612267。升级oracle客户端，版本升级到10.2.0.4(10.2.0.1之上就行)