| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> Re: startup/shutdown problem - long
Do you have some full directorys? If your archivelog directory it is full
for instance you will have problems. If you system tablespace is full and
you can't write to the dictionary it can hang. If the directory where one
of your control files is is full it will hang etc. Check for space
problems, in other words.
HTH,
Ruth
----- Original Message -----
From: Sindu <sindu_at_bigfoot.com>
To: Multiple recipients of list ORACLE-L <ORACLE-L_at_fatcity.com>
Sent: Friday, July 28, 2000 6:55 AM
Subject: startup/shutdown problem - long
Hi all,
Oracle 8.0.5.2.1, Solaris 2.6, 1GB Ram, 2 instances/2 databases...
I had problem shutting down both our db today, nothing strange in the alert file, no trace file, the alert is like this:
Thu Jul 27 14:29:59 2000
Shutting down instance (normal)
License high water mark = 2
Thu Jul 27 14:29:59 2000
ALTER DATABASE CLOSE NORMAL
Thu Jul 27 14:30:00 2000
SMON: disabling tx recovery
SMON: disabling cache recovery
Thu Jul 27 14:30:00 2000
Thread 1 closed at log sequence 1
Thu Jul 27 14:30:01 2000
Completed: ALTER DATABASE CLOSE NORMAL
Thu Jul 27 14:30:01 2000
ALTER DATABASE DISMOUNT
Completed: ALTER DATABASE DISMOUNT
And it just hangs there, there aren't any CPU / disk activity.. I waited almost 1 hour before deciding to shutdown abort.. And after shutdown abort, now I can't start (mount) the database, the alert showed:
PMON started with pid=2 DBW0 started with pid=3 LGWR started with pid=4 CKPT started with pid=5 SMON started with pid=6 RECO started with pid=7
and it hang here...
After that I tried to shutdown the other instance/database, and I had exactly same problem for 2nd instance...
I search metalink and found suggestions to reboot the machine, something wrong about the shared memory... Anyway this is our development machine so I have time to do some experiment with it, its uptime is already 267 days, so I'm trying to preserve it :)
I then did truss for svrmgrl, as suggested from metalink:
truss -fae -vall -o /tmp/mytruss svrmgrl
SVRMGR> connect internal;
Connected.
SVRMGR> startup nomount;
ORACLE instance started.
Total System Global Area 121265680 bytes Fixed Size 48656 bytes Variable Size 18735104 bytes Database Buffers 102400000 bytes Redo Buffers 81920 bytesSVRMGR> And I got the following repeated again and again non-stop:
26674: semop(1638400, 0xEFFFE9D0, 1) (sleeping...) 26678: semop(1638400, 0xEFFFE920, 1) (sleeping...) 26682: semop(1638400, 0xEFFFE920, 1) (sleeping...) 26686: semop(1638400, 0xEFFFE920, 1) (sleeping...) 26700: semop(1638400, 0xEFFFE920, 1) (sleeping...) 26674: Received signal #14, SIGALRM, in semop() [caught] 26674: semop(1638400, 0xEFFFE9D0, 1) Err#91 ERESTART 26674: semnum=2 semop=-1 semflg=0 26674: sigprocmask(SIG_BLOCK, 0xEFFFE5E8, 0x00000000) = 0 26674: set = 0 0 0 0 26674: times(0xEFFFE578) = -1987064759 26674: utim=1 stim=5 cutim=0 cstim=0 (HZ=100) 26674: setitimer(ITIMER_REAL, 0xEFFFE578, 0x00000000) = 0 26674: value: interval: 0.000000 sec value: 3.000000 sec 26674: sigprocmask(SIG_UNBLOCK, 0xEFFFE5E8, 0x00000000) = 0 26674: set = 0 0 0 0 26674: setcontext(0xEFFFE6B8) 26678: Received signal #14, SIGALRM, in semop() [caught] 26678: semop(1638400, 0xEFFFE920, 1) Err#91 ERESTART 26678: semnum=3 semop=-1 semflg=0 26678: sigprocmask(SIG_BLOCK, 0xEFFFE538, 0x00000000) = 0 26678: set = 0 0 0 0 26678: times(0xEFFFE4C8) = -1987064746 26678: utim=2 stim=1 cutim=0 cstim=0 (HZ=100) 26678: setitimer(ITIMER_REAL, 0xEFFFE4C8, 0x00000000) = 0 26678: value: interval: 0.000000 sec value: 3.000000 sec 26678: sigprocmask(SIG_UNBLOCK, 0xEFFFE538, 0x00000000) = 0 26678: set = 0 0 0 0 26678: setcontext(0xEFFFE608) 26682: Received signal #14, SIGALRM, in semop() [caught] 26682: semop(1638400, 0xEFFFE920, 1) Err#91 ERESTART 26682: semnum=4 semop=-1 semflg=0 26682: sigprocmask(SIG_BLOCK, 0xEFFFE538, 0x00000000) = 0 26682: set = 0 0 0 0 26682: times(0xEFFFE4C8) = -1987064733 26682: utim=1 stim=3 cutim=0 cstim=0 (HZ=100) 26682: setitimer(ITIMER_REAL, 0xEFFFE4C8, 0x00000000) = 0 26682: value: interval: 0.000000 sec value: 3.000000 sec 26682: sigprocmask(SIG_UNBLOCK, 0xEFFFE538, 0x00000000) = 0 26682: set = 0 0 0 0 26682: setcontext(0xEFFFE608) 26686: Received signal #14, SIGALRM, in semop() [caught] 26686: semop(1638400, 0xEFFFE920, 1) Err#91 ERESTART 26686: semnum=5 semop=-1 semflg=0 26686: sigprocmask(SIG_BLOCK, 0xEFFFE538, 0x00000000) = 0 26686: set = 0 0 0 0 26686: times(0xEFFFE4C8) = -1987064712 26686: utim=0 stim=6 cutim=0 cstim=0 (HZ=100) 26686: setitimer(ITIMER_REAL, 0xEFFFE4C8, 0x00000000) = 0 26686: value: interval: 0.000000 sec value: 3.000000 sec 26686: sigprocmask(SIG_UNBLOCK, 0xEFFFE538, 0x00000000) = 0 26686: set = 0 0 0 0 26686: setcontext(0xEFFFE608) 26700: Received signal #14, SIGALRM, in semop() [caught] 26700: semop(1638400, 0xEFFFE920, 1) Err#91 ERESTART 26700: semnum=7 semop=-1 semflg=0 26700: sigprocmask(SIG_BLOCK, 0xEFFFE538, 0x00000000) = 0 26700: set = 0 0 0 0 26700: times(0xEFFFE4C8) = -1987064676 26700: utim=2 stim=3 cutim=0 cstim=0 (HZ=100) 26700: setitimer(ITIMER_REAL, 0xEFFFE4C8, 0x00000000) = 0 26700: value: interval: 0.000000 sec value: 3.000000 sec 26700: sigprocmask(SIG_UNBLOCK, 0xEFFFE538, 0x00000000) = 0 26700: set = 0 0 0 0 26700: setcontext(0xEFFFE608)
$ps -fe|grep ora_
oracle 26682 1 0 16:27:08 ? 0:00 ora_lgwr_TEST oracle 26700 1 0 16:27:08 ? 0:00 ora_reco_TEST oracle 26696 1 0 16:27:08 ? 0:00 ora_smon_TEST oracle 26678 1 0 16:27:07 ? 0:00 ora_dbw0_TEST oracle 26686 1 0 16:27:08 ? 0:00 ora_ckpt_TEST oracle 26674 1 0 16:27:07 ? 0:00 ora_pmon_TEST
And when i type "alter database mount", it just hang and truss output is giving error on semop Err #91 ERESTART like above...
Anybody has any idea how to solve this (preferrably by not rebooting the sun
machine)??
I have tried removing (ipcrm) all shared memory/semaphore before running
svrmgrl, still no luck... Btw, isn't it strange that the key for the above
semaphore is 00000000 (?)
/etc/system:
set shmsys:shminfo_shmmni=200 set shmsys:shminfo_shmmax=4294967295 set shmsys:shminfo_shmseg=200 set semsys:seminfo_semmap=1000 set semsys:seminfo_semmsl=1000Received on Fri Jul 28 2000 - 10:15:42 CDT
 |
 |