Continuous backup of logical logs

Continuous backup of logical logs

Post by Pentti Tiet?v?ine » Sat, 11 May 2002 19:10:01



I'm using a script for unanttended logical logs backup (basically command
ontape -c &). The problem is that the background process seems to stop
(killed) almost once every day. Has anybody ideas for correcting the problem
or perhaps a better solution for backup of locigal logs?

Thanks in advance!
Pentti

 
 
 

Continuous backup of logical logs

Post by bick » Sat, 11 May 2002 20:08:55


Running ontape on background is'nt recommended and supported by
informix.
Ontape is an interactive tool, so if you would like to know what
happens, you could try starting it with nohup.

You can better use onbar.



Quote:>I'm using a script for unanttended logical logs backup (basically command
>ontape -c &). The problem is that the background process seems to stop
>(killed) almost once every day. Has anybody ideas for correcting the problem
>or perhaps a better solution for backup of locigal logs?

>Thanks in advance!
>Pentti


 
 
 

Continuous backup of logical logs

Post by Mark D. Stoc » Sun, 12 May 2002 08:20:44



> I'm using a script for unanttended logical logs backup (basically command
> ontape -c &). The problem is that the background process seems to stop
> (killed) almost once every day. Has anybody ideas for correcting the problem
> or perhaps a better solution for backup of locigal logs?

Run ontape -c in foreground.

Cheers,
--
Mark.

+----------------------------------------------------------+-----------+

| http://www.informix.com  http://www.informixhandbook.com |/////  / //|
| http://www.iiug.org  +-----------------------------------+////  / ///|
|                      |This email will self-destruct in   |///  / ////|
|                      |10 sec. If you received this email |//  / /////|
|                      |in error, sorry about the mess.    |/  ////////|
+----------------------+-----------------------------------+-----------+

 
 
 

Continuous backup of logical logs

Post by William Ri » Tue, 14 May 2002 22:09:11



> I'm using a script for unanttended logical logs backup (basically command
> ontape -c &). The problem is that the background process seems to stop
> (killed) almost once every day. Has anybody ideas for correcting the problem
> or perhaps a better solution for backup of locigal logs?

> Thanks in advance!
> Pentti

That is a pretty odd symptom.  Are you sure the box isn't getting
bounced every day, or anything weird like that?

Write a script wrapped around the ontape -c to capture it's output,
and return code when it exits.  Run that script in the background ...
Using the ontape output and return code, it shouldn't be to hard to
figure out what's going on.

Is there any chance the issue is you are running out of tapes, or are
you going to disk?

Anyway, ontape -c is a feasible solution, though you would be better
off using ontape -a out of the ALARMPROGRAM.  I believe iiug has some
sample ALARMPROGRAMS, but I am not sure on that...

Hope this helps,
Will

--
#!/usr/bin/ksh
ontape -c >logfile 2>&1
echo $? >>logfile

 
 
 

Continuous backup of logical logs

Post by Scott MacKenzi » Wed, 15 May 2002 00:23:01


Pentti,

It sounds like you may have a cron job set up that is looking for
zombie processes, or potentially "OLD" and non-root processes, to
be cleaned up for system optimization.



> > I'm using a script for unanttended logical logs backup
> > (basically command ontape -c &). The problem is that the
> > background process seems to stop (killed) almost once every
> > day. Has anybody ideas for correcting the problem or perhaps
> > a better solution for backup of locigal logs?

> > Thanks in advance!
> > Pentti

> That is a pretty odd symptom.  Are you sure the box isn't
> getting bounced every day, or anything weird like that?

> Write a script wrapped around the ontape -c to capture it's
> output, and return code when it exits.  Run that script in the
> background ...
> Using the ontape output and return code, it shouldn't be to hard
> to figure out what's going on.

> Is there any chance the issue is you are running out of tapes,
> or are you going to disk?

> Anyway, ontape -c is a feasible solution, though you would be
> better off using ontape -a out of the ALARMPROGRAM.  I believe
> iiug has some sample ALARMPROGRAMS, but I am not sure on that...

> Hope this helps,
> Will

> --
> #!/usr/bin/ksh
> ontape -c >logfile 2>&1
> echo $? >>logfile

--

   (                                               ______
   ))   .-- Scott MacKenzie; Dine' College IS --.   >===<--.
 C|~~| (>--- Phone/Voice Mail:  928-724-6639 ---<) |   = |-'

 
 
 

Continuous backup of logical logs

Post by Pentti Tiet?v?ine » Wed, 15 May 2002 19:54:53


Thank you your comments! My backup system is following:

Every day (5:00 am) cron starts a process which stops (kills) the ontape -c
process and runs the backup ontape -s to do 0-level backup. After that it
initializes the logbackupfile (copies /dev/null to the logbackupfile) and
starts the ontape -c command again. Logs are backed up to hard disk file. It
may sound a bit "old fashioned" but it has served us well.

I've configured the "logbackuptape" LTAPESIZE 4096000 and it should be able
to grow up to 4GB! The nohup.out file which seems to register ontape -c
commands output shows only "/usr/informix/logbackup.sh[8] 12345 Killed".
Somehow it points out that the operating system is killing this process, but
why it has not done it before (the system has been in use over one year!).

Thanks for any comments!

Pentti



Quote:> I'm using a script for unanttended logical logs backup (basically command
> ontape -c &). The problem is that the background process seems to stop
> (killed) almost once every day. Has anybody ideas for correcting the
problem
> or perhaps a better solution for backup of locigal logs?

> Thanks in advance!
> Pentti

 
 
 

Continuous backup of logical logs

Post by William Ri » Thu, 16 May 2002 00:02:01


Well the nohup would show the killed if you were using kill -9.

Are you sure your restart on ontape -c is working?

If you are backing up to disk, unless you are using the same file for
your level 0 backups, you should not have to stop the ontape -c
becausae of the backup.

IIRC, if you want to stop the ontape -c the proper thing to do is kill
-2 on the ontape process then do an onmode -l.  This simulates what
you would do on the console(which is hit cntrl-c and then when the
next log finishes ontape -c quits.)  One warning, it has been 2 years
since I worked with ontape so my memory is a little fuzzy.

Hope this helps,
Will


> Thank you your comments! My backup system is following:

> Every day (5:00 am) cron starts a process which stops (kills) the ontape -c
> process and runs the backup ontape -s to do 0-level backup. After that it
> initializes the logbackupfile (copies /dev/null to the logbackupfile) and
> starts the ontape -c command again. Logs are backed up to hard disk file. It
> may sound a bit "old fashioned" but it has served us well.

> I've configured the "logbackuptape" LTAPESIZE 4096000 and it should be able
> to grow up to 4GB! The nohup.out file which seems to register ontape -c
> commands output shows only "/usr/informix/logbackup.sh[8] 12345 Killed".
> Somehow it points out that the operating system is killing this process, but
> why it has not done it before (the system has been in use over one year!).

> Thanks for any comments!

> Pentti



> > I'm using a script for unanttended logical logs backup (basically command
> > ontape -c &). The problem is that the background process seems to stop
> > (killed) almost once every day. Has anybody ideas for correcting the
>  problem
> > or perhaps a better solution for backup of locigal logs?

> > Thanks in advance!
> > Pentti

 
 
 

Continuous backup of logical logs

Post by Rich Bl » Thu, 16 May 2002 00:04:39


The mentioning of the LTAPESIZE parameter has reminded me of a problem
I had with ontape. On a lot of versions of IDS (you don't mention what
version or OS you're using) ontape cannot write to files greater than
2Gb, even if the OS has large file support. This is something to do
with the limitations of seek() that it uses.

Could it be that your log backups have started getting bigger than 2Gb
a day? If so ontape will exit with an error code if the file gets too
big.

Cheers

Richb


> Thank you your comments! My backup system is following:

> Every day (5:00 am) cron starts a process which stops (kills) the ontape -c
> process and runs the backup ontape -s to do 0-level backup. After that it
> initializes the logbackupfile (copies /dev/null to the logbackupfile) and
> starts the ontape -c command again. Logs are backed up to hard disk file. It
> may sound a bit "old fashioned" but it has served us well.

> I've configured the "logbackuptape" LTAPESIZE 4096000 and it should be able
> to grow up to 4GB! The nohup.out file which seems to register ontape -c
> commands output shows only "/usr/informix/logbackup.sh[8] 12345 Killed".
> Somehow it points out that the operating system is killing this process, but
> why it has not done it before (the system has been in use over one year!).

> Thanks for any comments!

> Pentti



> > I'm using a script for unanttended logical logs backup (basically command
> > ontape -c &). The problem is that the background process seems to stop
> > (killed) almost once every day. Has anybody ideas for correcting the
>  problem
> > or perhaps a better solution for backup of locigal logs?

> > Thanks in advance!
> > Pentti

 
 
 

Continuous backup of logical logs

Post by Mark D. Stoc » Thu, 16 May 2002 01:22:53



> Thank you your comments! My backup system is following:

> Every day (5:00 am) cron starts a process which stops (kills) the ontape -c
> process and runs the backup ontape -s to do 0-level backup. After that it
> initializes the logbackupfile (copies /dev/null to the logbackupfile) and
> starts the ontape -c command again. Logs are backed up to hard disk file. It
> may sound a bit "old fashioned" but it has served us well.

As long as you are backing up your logs to tape at some point.

Quote:> I've configured the "logbackuptape" LTAPESIZE 4096000 and it should be able
> to grow up to 4GB! The nohup.out file which seems to register ontape -c
> commands output shows only "/usr/informix/logbackup.sh[8] 12345 Killed".
> Somehow it points out that the operating system is killing this process, but
> why it has not done it before (the system has been in use over one year!).

> Thanks for any comments!

Maybe you have hit a 2Gb file size limit in the OS?

Cheers,
--
Mark.

+----------------------------------------------------------+-----------+

| http://www.informix.com  http://www.informixhandbook.com |/////  / //|
| http://www.iiug.org  +-----------------------------------+////  / ///|
|                      |This email will self-destruct in   |///  / ////|
|                      |10 sec. If you received this email |//  / /////|
|                      |in error, sorry about the mess.    |/  ////////|
+----------------------+-----------------------------------+-----------+

 
 
 

Continuous backup of logical logs

Post by Murray Woo » Thu, 16 May 2002 07:22:06


Do not run ontape -c in background.  You could use ALARMPROG to backup each
log as it completes.  You also do not need to stop log backup to do an
archive.

MW


> Thank you your comments! My backup system is following:

> Every day (5:00 am) cron starts a process which stops (kills) the
ontape -c
> process and runs the backup ontape -s to do 0-level backup. After that it
> initializes the logbackupfile (copies /dev/null to the logbackupfile) and
> starts the ontape -c command again. Logs are backed up to hard disk file.
It
> may sound a bit "old fashioned" but it has served us well.

> I've configured the "logbackuptape" LTAPESIZE 4096000 and it should be
able
> to grow up to 4GB! The nohup.out file which seems to register ontape -c
> commands output shows only "/usr/informix/logbackup.sh[8] 12345 Killed".
> Somehow it points out that the operating system is killing this process,
but
> why it has not done it before (the system has been in use over one year!).

> Thanks for any comments!

> Pentti



> > I'm using a script for unanttended logical logs backup (basically
command
> > ontape -c &). The problem is that the background process seems to stop
> > (killed) almost once every day. Has anybody ideas for correcting the
>  problem
> > or perhaps a better solution for backup of locigal logs?

> > Thanks in advance!
> > Pentti

 
 
 

1. remote db restore continuous logical log recovery

I have a customer that wants to do the following:

1. they have a primary db on a box in atlanta
2. they have a secondary (drp) db on a box in washington
3. they want to take a level 0 ontape (could be onbar) of primary
4. send this level 0 to washington and restore secondary from level 0
5. Then from this point on they want to send the logical logs to
washington and continually keep the secondary db up to date with in 8
hours of the primary.

I have mentioned to them that they would be better off with Data
Replication. However, they are not comfortable with that. Do you know
if what they have proposed above will work?

If you can point me to some documentation, let me know and I will look
it up myself.

Thanks,
Gary

2. Smart

3. [Fwd: remote db restore continuous logical log recovery]

4. SP referencing objects across databases

5. Logical Log Continous Backup - ArcServer Software Backup

6. FmPro - Director

7. Continuous log backup in the background

8. HELP ODBC DRIVER PERFOMANCE !SUPER URGENT

9. Continuous Log Backup

10. continuous log backup to disk (ontape -c) ...

11. continuous log backup to disk (ontape -c) ... again

12. Continuous Log Backup to Disk

13. deciding on logical log backup strategy - advice, anyone?