Printable View
hi, in svrmgrl, i did a shutdown, the system replied that the database is shutdown. however, i did a ps -ef | grep ora, and i still find these oracle processes. afterwhich i did a startup, and the system told me that oracle is still running. what is wrong?
you didnt actually shut it down, lets see what you typed and variables and such like
You can try shutting down the database again, then kill the othe oracle process manually, then you can start your db.
Can you paste output of the command
ps -ef | grep ora
Do you have multiple databases on the system? If so, did you set the ORACLE_SID to shutdown the appropriate database? The SID name is appended to oracle processes.
I guess your listener service is still alive.
In a normal shutdown, oracle will wait untill all the connections are closed and then only it will shutdown.
You might want to issue shutdown immediate/shutdown abort
Sam
i wished i could. but now the database is okay and running. anyway, the output of ps -ef | grep ora is the oracle processes with smon, pmon, archiver, lgwr, dbwr and etc.......Quote:
Can you paste output of the command
ps -ef | grep ora
this is what actually happensQuote:
Originally posted by sambavan
In a normal shutdown, oracle will wait untill all the connections are closed and then only it will shutdown.
You might want to issue shutdown immediate/shutdown abort
Sam
1) we had scripts to shutdown/startup the database for offline backup in our crontab entries
2) and these scripts had been working fine.
3) but recently, there had been 2 times when the scripts are not working
4) scenario: couldnt startup database therefore go and analyze
5) found that the shutdown script did not shutdown the database. Cos going into svrmgrl, and doing a startup says that "the oracle instance is still running"
5.1) do a ps -ef | grep ora and indeed saw all these oracle processes including the shutdown script entry.
5.2) going back to svrmgrl, did a shutdown immediate. still same result in 5.1
5.3) colleague did a shutdown abort and it works.
my questions
1)is that isnt it dangerous to do a shutdown abort. actually, i was thinking about killing all the orace processes. so which is more advisable?
2) what are the valid reasons doing a shutdown abort.
thanks
Shutdown abort and killing the background process are same and both are not advisable.Quote:
Originally posted by yls177
my questions
1)is that isnt it dangerous to do a shutdown abort. actually, i was thinking about killing all the orace processes. so which is more advisable?
Sometimes it happens that when long batch processes or long transactions are running and you want to take backups and do a shutdown immediate then it might take hours for a shutdown or even look as if there is no response. Then you can possible do a shutdown abort. But be mentally prepared for unexpected results :). And yes after a shutdown abort do not forget to startup and do a shutdown immediate again before taking the backup.Quote:
Originally posted by yls177
2) what are the valid reasons doing a shutdown abort.
thanks
Other reasons can be....
1. When the building is on fire and you need to run :).
2. When you need to show the developers who is in control.
3. When you want to act coooool and show off that you can recover from a shutdown abort.
4. When you need to .......................
Contd. on the last page...
where is the last page?Quote:
Originally posted by adewri
.............
4. When you need to .......................
Contd. on the last page... [/B]
hi, below are my steps
1) shutdown immediate;
2) but oracle still hanging, as in can do a ps -ef | grep ora and still the set of family oracle processes are alive
3) so i do a select * from v$instance;
4) and when it says that " oracle not available"
5) then i safely do a shutdown abort
6) finally, the family set of oracle processes are gone.
but is it absolutely safe to do this?
please confirm
If it's the only database running on that machine, run ipcs to see if the Shared Memory Segments and Semaphore Arrays are still alive. It's rare, but some times they don't die if you've had an abnormal database abortion.
ipcrm will remove them (check man pages for syntax). This WILL definately shut your database down. It's as effective as a power switch on a unix/linux operating system.
Best of luck.
shutdown abort is a last resort and ipcrm is a final and fatal resort.
You shouldn't have any trouble bring the database back up. It'll do a delayed recovery on it's own when it comes back up.
in my alert file, i had the following entries.
Thu Nov 20 22:05:29 2003
Shutting down instance (immediate)
License high water mark = 122
Thu Nov 20 22:08:06 2003
Shutting down instance (abort)
License high water mark = 122
Instance terminated by USER, pid = 115688
Thu Nov 20 22:08:16 2003
alter database open
Beginning crash recovery of 1 threads
Thu Nov 20 22:08:17 2003
Thread recovery: start rolling forward thread 1
Recovery of Online Redo Log: Thread 1 Group 1 Seq 244585 Reading mem 0
Mem# 0 errs 0: /oracle/SID/origlogA/log_g15m1.dbf
Mem# 1 errs 0: /oracle/SID/mirrlogA/log_g15m2.dbf
Thu Nov 20 22:08:17 2003
Thread recovery: finish rolling forward thread 1
Thread recovery: 3 data blocks read, 3 data blocks written, 2 redo blocks read
Crash recovery completed successfully
my questions
1) both shutdown statements (one is immediate, the other is abort). does that mean that there are 2 shutdown statements, one is shutdown immediate, and the other is shutdown abort.
2) also, oracle did recovered the database successfully?
thanks
Looks like you did a shutdown immediate, then abort, then started up.
It also looks like the startup recovery was good.
okay.. i guess i am fortunate this time. cos i think shutdown abort will gives some database big headaches like inconsistency, database cant startup and etc.. am i right
shutdown abort isn't a bad thing, only in rare cases will a shutdown abort give you problems where you lose an online redo log.
Normally Oracle is very good at recovering itself from that so dont worry too much if you have to do one
i am quite a cautious guy.... so wont try dangerous stuffs... ;) i am concerned cos my colleagues kept doing shutdown abort when immediate takes a while....
i told them things like rollback, switch logfile and etc./..
we do shutdown abort all the time
part of backup process when taking cold backups
hi pando,
do you take the cold backup just after a shutdown abort.
is it OK.
Raja
we do
alter system switch logfile
alter system checkpint
shutdowna bort
startup restrict
shutdown immediate
We do exactly the same as a standard on 100+ prod databases - had very little problem with it.
Quote:
Originally posted by pando
we do
alter system switch logfile
alter system checkpint
my understanding is taht alter system switch logfile triggers a checkpoint. and pando did this so as to update the changes to the redo log files right?
here is where i dont understand.. after both alter statements in the earlier quote, why need the above three statements. wont a shutdown immediate do the trick...Quote:
shutdowna bort
startup restrict
shutdown immediate
i will do the below (from pando ones...)
alter system switch logfile
alter ssytem checkpoint
shutdown immediate
is there a difference?
shutdown imemdiate executes a checkpoint implicitly
we do shutdown abort because sometimes it tkes 10 minutes to shutdown immediate a database
shutdown abort hmm takes 1 second
okay.. another case of shutdown immediate...
waited for more than 10mins..
then did a shutdown abort...
oracle displays this "ORA-01089: immediate shutdown in progress - no operations are permitted"
and in my alert file, i had this "License high water mark = 2" which indicates that there are only 2 sessions only, right?
also from this http://metalink.oracle.com/metalink/...&p_id=164504.1
i think thats all... to my queries.. thanks folks for bearing with me