求助，informix 自己down了。

test_123

最近informix 自己down了好几次，不知道什么原由，Informix已经用了好几年都没问题了，最近不知道是怎么回事。

test_123

13:01:36  Checkpoint Completed:  duration was 0 seconds.
13:01:36  Checkpoint loguniq 585, logpos 0x602018
13:04:36  Checkpoint Completed:  duration was 0 seconds.
13:04:36  Checkpoint loguniq 585, logpos 0x603018
13:05:18  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:05:18  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:05:46  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:05:46  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:07:38  Checkpoint Completed:  duration was 2 seconds.
13:07:38  Checkpoint loguniq 585, logpos 0x604018
13:08:35  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:08:35  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:08:35  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:08:35  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:08:35  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:08:35  shmget: [EEXIST][17]: key 52564804: shared memory already exists

test_123

Informix Dynamic Server 2000 Version 9.21.UC2

13:14:48  Assert Failed: No Exception Handler
13:14:48   Who: Session(1358, crol@a5d0a2b, -1, 0)
                Thread(9, soctcplst, 0, 1)
                File: mtex.c Line: 405
13:14:48   Results: Exception Caught. Type: MT_EX_OS, Context: mem
13:14:48   Action: Please notify Informix Technical Support.
13:14:48  Stack for thread: 9 soctcplst

base: 0x11747018
  len:   33280
   pc: 0x08511aa0
  tos: 0x1174de68
state: running
   vp: 1

===========------------- - - - - - -
tail -100 /opt/informix/online.log:


13:14:42  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:14:42  out of virtual shared memory

13:14:42  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:14:42  out of virtual shared memory

13:14:42  shmget: [EEXIST][17]: key 52564804: shared memory already exists

13:14:48  shmget: [EEXIST][17]: key 52564804: shared memory already exists
13:14:48  out of virtual shared memory

13:14:48  Assert Failed: No Exception Handler
13:14:48  Informix Dynamic Server 2000 Version 9.21.UC2
13:14:48   Who: Session(1358, crol@a5d0a2b, -1, 0)
Thread(9, soctcplst, 0, 1)
File: mtex.c Line: 405
13:14:48   Results: Exception Caught. Type: MT_EX_OS, Context: mem
13:14:48   Action: Please notify Informix Technical Support.
13:14:48  stack trace for pid 32221 written to /tmp/af.3f19447
13:14:48   See Also: /tmp/af.3f19447, shmem.3f19447.0


===========------------- - - - - - -

test_123

之前还出过其他错误，
7:22:15
07:22:15  Informix Dynamic Server 2000 Version 9.21.UC2     Software Serial Number AAC#J938587

07:22:15  Assert Failed: Assert Failed: %s
07:22:15   Who: Session(0, @, 0, 0)
                Thread(4, aio vp 0, 0, 5)
                File: mt.c Line: 7218
07:22:15  Stack for thread: 4 aio vp 0

base: 0x116cb018
  len:   33280
   pc: 0x08511aa0
  tos: 0x116d2bbc
state: running
   vp: 5

test_123

02:57:36  Checkpoint Completed:  duration was 0 seconds.
02:57:36  Checkpoint loguniq 582, logpos 0x5ea018

02:59:12  Dynamically allocated new virtual shared memory segment (size 40960KB)
03:00:36  Checkpoint Completed:  duration was 0 seconds.
03:00:36  Checkpoint loguniq 582, logpos 0x5eb018

03:01:05  Dynamically allocated new virtual shared memory segment (size 40960KB)
03:03:00  Dynamically allocated new virtual shared memory segment (size 40960KB)
03:03:36  Checkpoint Completed:  duration was 0 seconds.
03:03:36  Checkpoint loguniq 582, logpos 0x5ec018

03:04:53  Dynamically allocated new virtual shared memory segment (size 40960KB)
03:06:36  Checkpoint Completed:  duration was 0 seconds.
03:06:36  Checkpoint loguniq 582, logpos 0x5ed018

test_123

开始自己觉得是shm的错，就修改了
onconfig里的
SHMVIRTSIZE     40960           # initial virtual shared memory segment size
SHMADD          40960           # Size of new shared memory segments (Kbytes)
SHMTOTAL        4096000         # Total shared memory (Kbytes). 0=>unlimited
CKPTINTVL       180             # Check point interval (in sec)
加大了1倍。
CKPTINTVL加到了5分钟

test_123

怪啊。。。。。。等待中。。。。。。

wolfop

没什么好奇怪的9.21有几个BUG没有补丁，必须升级到9.4以后版本。

liaosnet

日志不全....

hakajawa

SHMVIRTSIZE     40960           # initial virtual shared memory segment size
改大点