Shulou(Shulou.com)06/01 Report--

The database is 11g physical DG

Standby database reported error Process M000 died, see its trace file.

Now the error of online search is that the resource is exhausted and cannot be connected, but the database is still alive.

This situation is solved by adjusting the number of resources, such as processes reaching the upper limit.

View resource usage

Select * from v$resource_limit RESOURCE_NAME CURRENT_UTILIZATION MAX_UTILIZATION INITIAL_ALLOCATION LIMIT_VALUE -processes 50 54 150 150sessions 64 68 252 252enqueue_locks 35 39 3310 3310enqueue_resources 17 17 1328 UNLIMITEDges_procs 0 0 0 0ges_ress 0 0 0 UNLIMITEDges_locks 0 0 0 UNLIMITEDges_cache_ress 0 0 0 UNLIMITEDges_reg_msgs 0 0 0 UNLIMITEDges_big_msgs 0 0 0 UNLIMITEDges_rsv_msgs 0 0 0 0gcs_resources 0 0 0 0gcs_shadows 0 0 0 0dml_locks 0 0 1108 UNLIMITEDtemporary_table_locks 0 0 UNLIMITED UNLIMITEDtransactions 5 5 277 UNLIMITEDbranches 0 0 277 UNLIMITEDcmtcallbk 2 3 277 UNLIMITEDmax_rollback_segments 11 11 277 65535sort_segment_locks 0 1 UNLIMITED UNLIMITEDk2q_locks 0 0 504 UNLIMITEDmax_shared_servers 1 1 UNLIMITED UNLIMITEDparallel_max_servers 00 135 3600

Add processes

Alter system set processes=200 scope=spfile

Then restart the database to solve.

But my problem cannot be solved by the above methods.

Observe the problem example again and find the following:

Login through sqlplus, the entry is the idle process, can only kill the os process to restart.

It shuts down automatically after rebooting.

Check the alert log and have no information

Use the strace command to trace the smon process display

[oracle@rac2 ~] $strace-p 7351Process 7351 attachedgetrusage (RUSAGE_SELF, {ru_utime= {0, 21996}, ru_stime= {0, 14997},...}) = 0getrusage (RUSAGE_SELF, {ru_utime= {0, 21996}, ru_stime= {0, 14997},...}) = 0semtimedop (15794179, {{18,-1, 0}, 1, {3, 0}) =-1 EAGAIN (Resource temporarily unavailable) semtimedop (15794179, {18,-1, 0}, 1 {3,0}) =-1 EAGAIN (Resource temporarily unavailable) semtimedop (15794179, {{18,-1,0}}, 1, {3,0}) =-1 EAGAIN (Resource temporarily unavailable) semtimedop (15794179, {{18,-1,0}}, 1, {3,0}) =-1 EAGAIN (Resource temporarily unavailable)

You can see that the resource is no longer available, and it is suspected that it is a memory problem.

It is explained here that the server is used for testing, and there are several instances running above, such as 5 instances such as single node 10g force 11g 12c standby,11g rac and so on.

Use ipcs-m to view

Ipcs-m-Shared Memory Segments-key shmid owner perms bytes nattch status 0x1644d05c 28082176 oracle 945815552 28 0xb4a8f6f0 28147713 oracle 660 4096 0 0xb3928380 28475394 oracle 640 14680064 82 0x00000000 28508163 oracle 640 868220928 82 0xb2f44a00 41123844 oracle 660 4096 0 0x27bbfbde 3309577 oracle 755 1079228 1 0x00000000 22708244 oracle 640 33554432 49 0x00000000 22741013 oracle 640 2449473536 49 0x0fc14ec0 22773782 oracle 640 2097152 49

You can see that there are 6 semaphores above, excluding 0x00000000 (this semaphore is derived)

However, there are a total of 5 instances. It is suspected that the memory segment of standby is abnormally closed, and the system did not clean it up.

And there is an oracle section with permission of 755, which is usually 640, which is suspected to be caused by this, and

To see which instance of oracle each memory segment belongs to, you can use the oracle command sysresv

[oracle@rac2 ~] $sysresvIPC Resources for ORACLE_SID "orcl": Shared Memory:ID KEY41385984 0xb2f44a00Semaphores:ID KEY16220162 0xb38b1d5cOracle Instance alive for sid "orcl"

Multiple instances in the same version can be specified.

Sysresv-l sid

Delete shared memory segment

[oracle@rac2 trace] $ipcrm-- hlepusage: ipcrm [[- Q msqid] [- m shmid] [- s semid] [- Q msgkey] [- M shmkey] [- S semkey].]

Or sysresv-f sid delete

Restart stanby after deletion and everything is fine.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.