-
I have an oracle 8.1.6 database with archive log mode enabled and configured. I went into DBA studio and removed archive log mode and hit the apply button. DBA studio then attempts to bounce my database but has been on the same screen for about two hours, stuck at the "shutting down database" step. I don't want to end task because I will leave the database in an unknown state. And I don't have a dump file to restore from. How do I know if this operation is going to go through or not? What do I do if it just sits here?
-
What's the database doing? Are the disks spinning? Is there an error in the alert.log? Are any oracle processes using CPU?
-
The database doesn't appear to be doing anything... The RAID array isn't doing anything either... the image with the gears is still going though and I can move the box around... Task manager reports that the application is running but all the cpu cycles are dedicated to idle processes.
-
I would guess it is hung. To take your DB out of archivelog mode:
svrmgrl
connect internal
shutdown immediate
startup mount exclusive
alter database noarchivelog
shutdown immediate
startup
-
here is the error I got in the windows 2000 application log
Event Type: Error
Event Source: Oracle.ewpd
Event Category: None
Event ID: 20
Date: 2/1/2002
Time: 10:53:45 AM
User: N/A
Computer: EWPD01
Description:
Archive process error: ORACLE Instance ewpd - Can not allocate log, archival required
.
-
I can't log into server manager because it says that my database is in the process of shutting down.
-
Check your log_archive_dest, it may be full. Also, make sure you have automatic archiving turned on by checking the init.ora parameter log_archive_start=true.
-
I know I didn't have auto archiving turned on and since DBA studio is still doing something I can't make any changes at all. What would happen if I end tasked on DBA studio and restarted the server. Could I get my data back?
-
Same thing happened to me while I am using DBA studio for shutdown immediate , but not alter database noarchlog ;
That time I loggged as internal into svrmgrl .. I issued shutdown abort .. I know I don't have any transactions pending .. In your case if any transactions pending , they won't rollback .. It should allow you to logon as internal saying connected idle instance or some message ...
-
Thanks again Jeff... that's two in one day... lol
I did an end task and then a shutdown abort, removed archivelog and then restarted the db... everything seems to be working fine. Is there anything else I need to do now that the db is back up?
-
No, as long as you are in noarchivelog mode, you should be OK on that front.
-
Yeah everything seems to be running great. Have a peaceful weekend Jeff. Thanks again for everyones comments.
-
Of course, you understand the implications of not running in archivelog mode, right?
-
Implications of not running in archivelog mode? I'm not sure I understand, I am under the impression this is the preferable scenario. Could you explain?
-
-
How offen that you do a backup. If your database were to go down, or fail for some reason, then you would only be able to recover the data to the last backup. But with the archive logs, you can recover the database to the point of failure.
Sam
-
For the vast majority of our clients we set up AT scripts to run a bat file that does an export/warm backup of the database at certain off-peak times during the day. This occurs 3 times a day for most clients.
-
Here is an example of AT scripts that run once a day.
AT 6:00 /EVERY:M CMD /C C:\DATDUMP MONDAY
AT 6:00 /EVERY:T CMD /C C:\DATDUMP TUESDAY
AT 6:00 /EVERY:W CMD /C C:\DATDUMP WEDNESDAY
AT 6:00 /EVERY:Th CMD /C C:\DATDUMP THURSDAY
AT 6:00 /EVERY:F CMD /C C:\DATDUMP FRIDAY
AT 6:00 /EVERY:Sa CMD /C C:\DATDUMP SATURDAY
AT 6:00 /EVERY:Su CMD /C C:\DATDUMP SUNDAY
The DATDUMP file looks like this:
EXP DATABASEUSERNAME/PASSWORD FILE=C:\DATADMPS\%1
-
So you can afford at least 8-hours worth of data loss? Good for you.
BTW, exp/imp is *no kind of database backup* method, I hope you are aware of that. And I hope you are aware that you are actually exporting the data is inconsistent (regarding one table content to another), because you are not using CONSISTENT=Y during your exp. And if your exp batch file realy consists only that line that you have written I realy do hope you are aware that you are exporting only one users' objects all the time.....
If your database encounters media failure you'll certanly have a lot of work to do....
