IDS 9.21
Linux 7.3
Ситуация такая
Упал Informix видимо из-за проблем со SCSI

Sep 23 21:11:44 tproliant1 kernel: scsi : aborting command due to timeout : pid 395156202, scsi0, ch
annel 0, id 2, lun 2 Write (10) 40 01 9e a4 48 00 03 f8 00
Sep 23 21:11:44 tproliant1 kernel: sym53c8xx_abort: pid=395156202 serial_number=395155344 serial_num
ber_at_timeout=395155344
Sep 23 21:11:44 tproliant1 kernel: sym53c896-0-<2,*>: control msgout: 82 20 1f d.
Sep 23 21:11:44 tproliant1 kernel: SCSI disk error : host 0 channel 0 id 2 lun 2 return code = 25050
07f
Sep 23 21:11:44 tproliant1 kernel: I/O error: dev 08:1d, sector 559136
Sep 23 21:11:44 tproliant1 kernel: SCSI disk error : host 0 channel 0 id 2 lun 2 return code = 25050
07f
Sep 23 21:11:44 tproliant1 kernel: I/O error: dev 08:17, sector 320191
Sep 23 21:11:44 tproliant1 kernel: SCSI disk error : host 0 channel 0 id 2 lun 2 return code = 25050
07f
Sep 23 21:11:44 tproliant1 kernel: I/O error: dev 08:17, sector 320064
Sep 23 21:11:44 tproliant1 kernel: SCSI disk error : host 0 channel 0 id 2 lun 2 return code = 25050
07f
Sep 23 21:11:44 tproliant1 kernel: I/O error: dev 08:16, sector 229036
Sep 23 21:11:44 tproliant1 kernel: SCSI disk error : host 0 channel 0 id 2 lun 2 return code = 25050
07f
Sep 23 21:11:44 tproliant1 kernel: I/O error: dev 08:1b, sector 2051384
Sep 23 21:11:44 tproliant1 kernel: sym53c896-0:2: message d sent on bad reselection.
Sep 23 21:19:21 tproliant1 shutdown: shutting down for system reboot
Sep 23 21:19:22 tproliant1 init: Switching to runlevel: 6



После временного устранения INformix перестал загружаться:


21:15:02 Event alarms enabled. ALARMPROG = '/usr/informix/etc/log_full.sh'
21:15:03 Booting Language <c> from module <>
21:15:03 Loading Module <CNULL>
21:15:03 Booting Language <builtin> from module <>
21:15:03 Loading Module <BUILTINNULL>
21:15:10 Informix Dynamic Server 2000 Version 9.21.UC3 Software Serial Number AAC#J892900
21:15:10 Informix Dynamic Server 2000 Initialized -- Shared Memory Initialized.

21:15:10 Physical Recovery Started.
21:15:12 Physical Recovery Complete: 362 Pages Restored.

21:15:12 Logical Recovery Started.
21:15:12 10 recovery worker threads will be started.
21:15:13 Fast Recovery Switching to Log 124322
21:15:13 Fast Recovery Switching to Log 124323
21:15:14 Fast Recovery Switching to Log 124324
21:15:14 Fast Recovery Switching to Log 124325
21:15:16 Logical Recovery Complete.
10911 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks

21:15:17 Dataskip is now OFF for all dbspaces
21:15:18 Checkpoint Completed: duration was 0 seconds.
21:15:18 Checkpoint loguniq 124325, logpos 0xce018

21:15:18 Dropping temporary TBLspace 100060, recovering 8 pages.
21:15:18 Dropping temporary TBLspace 100061, recovering 8 pages.
21:15:18 Dropping temporary TBLspace 100062, recovering 8 pages.
21:15:18 Dropping temporary TBLspace 100063, recovering 8 pages.
21:15:18 Dropping temporary TBLspace 100064, recovering 8 pages.
21:15:18 Dropping temporary TBLspace 100065, recovering 8 pages.
21:15:18 Dropping temporary TBLspace 100066, recovering 8 pages.
21:15:18 Dropping temporary TBLspace 100067, recovering 8 pages.
21:15:18 On-Line Mode
21:15:18 Informix Dynamic Server 2000 Stopped.

21:28:01 Informix Dynamic Server 2000 Started.
21:28:03 Segment locked: addr=0x10000000, size=418852864
Fri Sep 23 21:28:04 2005

21:28:04 Event alarms enabled. ALARMPROG = '/usr/informix/etc/log_full.sh'
21:28:04 Booting Language <c> from module <>
21:28:04 Loading Module <CNULL>
21:28:04 Booting Language <builtin> from module <>
21:28:04 Loading Module <BUILTINNULL>
21:28:12 Informix Dynamic Server 2000 Version 9.21.UC3 Software Serial Number AAC#J89290
21:28:12 Informix Dynamic Server 2000 Initialized -- Shared Memory Initialized.

21:28:12 Physical Recovery Started.
21:28:12 Physical Recovery Complete: 0 Pages Restored.

21:28:12 Logical Recovery Started.
21:28:12 10 recovery worker threads will be started.
21:28:12 Size of resident + virtual segments 409036KB + 934148KB > 972800KB
21:28:12 total allowed by configuration parameter SHMTOTAL
21:28:12 out of virtual shared memory

21:29:17 Cannot Rollforward from Checkpoint.
21:29:18 Informix Dynamic Server 2000 Stopped.

21:29:19 mt_shm_remove: WARNING: may not have removed all/correct segments
21:29:52 Informix Dynamic Server 2000 Started.
21:29:54 Segment locked: addr=0x10000000, size=418852864

-----
SHMTOTAL поменял на 0 (Unlimited)
IDS запустился , но добавил 4 виртуальных сегмента

------ Shared Memory Segments --------
key shmid owner perms bytes nattch status
0x52564801 229377 root 660 418852864 29 locked
0x52564802 262146 root 660 131072000 29
0x52564803 294915 root 660 268435456 29
0x52564804 327684 root 660 268435456 29
0x52564805 360453 root 660 268435456 29
0x52564806 393222 root 660 20189184 29

почему IDS решил добавить виртуальные сегменты?

Segment Summary:
id key addr size ovhd class blkused blkfree
229377 1381386241 10000000 418852864 227552 R* 102232 27
262146 1381386242 28f73000 131072000 4600 V 7349 24651
294915 1381386243 42135000 825495552 25800 V 7 201530
Total: - - 1375420416 - - 109588 226208

Клянусь, со значением SHMTOTAL = 972800, IDS перегружался много раз.
Вчера перезапустил IDS еще раз.
Стало как раньше:
------ Shared Memory Segments --------
key shmid owner perms bytes nattch status
0x52564801 32769 root 660 418852864 29 locked
0x52564802 65538 root 660 131072000 29

Может от проблем со SCSI информикс кешировал информацию в логах,
а при запуске - пытался ее выгрузить в shared memory?