IDS9.4FC8 backup problem
A user running ontape -s -L 0 from cron on IDS 9.4.FC8 under AIX found that some nights the level-0 archive never printed its usual "x percent done" progress messages, ran far longer than normal with heavy disk read/write, and logged no errors. John Miller suggested the known "very old page" timestamp-refresh behaviour during level-0 archives; the poster instead found APAR IC54415, and setting CCFLAGS made no difference. Keith Simmons argued the very-old-page bug only applied before 7.31xC8 and was fixed in code, and asked about tape connection, ONCONFIG tape settings and onstat -g arc output (also noting 8-9 hours for 300GB is slow). The thread ends with the user's TAPEBLK/TAPEDEV values and an unanswered question about CCFLAGS; no resolution is recorded.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
Dear all
I got a strange problem about using ontape (on IDS 9.4.FC8 with AIX).
I wrote a backup script like :
TAPE=`onstat -c | grep "^TAPEDEV" | awk '{print $2}'`
echo "Begin ontape backup to $TAPE: `date`"
ontape -s -L 0 <<!
! echo "Finish ontape backup to $TAPE: `date`"
I use it in the cron job to perform automatic backup and it workes fine all
the time. Usually it will finish before 6 am. However this morning it is still
runing till 9 am.In the log message, I usually can see the following message.
Please mount tape 1 on /IDS9.40/informix/tape and press Return to continue ...
10 percent done.
20 percent done.
30 percent done.
40 percent done.
50 percent done.
100 percent done.
Please label this tape as number 1 in the arc tape sequence.
This tape contains the following logical logs:
956
Program over
But today I didn't see those information in the log and there is no error
message in online.log. Besides that I saw that disk had high read/write.
Because it will make a low performance, I broke the program by manual.
Could anyone can help me to solve this problem?
thanks in advance
This sounds like the very old page issue. There is many post about it,=
but
in short every so often the backup has to update the timestamp of pages=
which
are not accessed for update. This is done by a level 0 archive and can=
general
be viewed by high read and write counts in an onstat -u.
The frequency you will see this depends on you update activity. You co=
uld
see it
today then not for another year. This issue has been removed in new
versions.
John F. Miller III
STSM, Support Architect
miller3@us.ibm.com
503-578-5645
IBM Informix Dynamic Server (IDS)
|------------>
| From: |
|------------>
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|"FAN YANG" <fania.yang@gmail.com> =
=
|
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|------------>
| To: |
|------------>
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|ids@iiug.org =
=
|
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|------------>
| Date: |
|------------>
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|01/21/2010 08:28 PM =
=
|
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|------------>
| Subject: |
|------------>
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|IDS9.4FC8 backup problem [18741] =
=
|
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|------------>
| Sent by: |
|------------>
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|ids-bounces@iiug.org =
=
|
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
Dear all
I got a strange problem about using ontape (on IDS 9.4.FC8 with AIX).
I wrote a backup script like :
TAPE=3D`onstat -c | grep "^TAPEDEV" | awk '{print $2}'`
echo "Begin ontape backup to $TAPE: `date`"
ontape -s -L 0 <<!
! echo "Finish ontape backup to $TAPE: `date`"
I use it in the cron job to perform automatic backup and it workes fine=
all
the time. Usually it will finish before 6 am. However this morning it i=
s
still
runing till 9 am.In the log message, I usually can see the following
message.
Please mount tape 1 on /IDS9.40/informix/tape and press Return to
continue ...
10 percent done.
20 percent done.
30 percent done.
40 percent done.
50 percent done.
100 percent done.
Please label this tape as number 1 in the arc tape sequence.
This tape contains the following logical logs:
956
Program over
But today I didn't see those information in the log and there is no err=
or
message in online.log. Besides that I saw that disk had high read/write=
.
Because it will make a low performance, I broke the program by manual.
Could anyone can help me to solve this problem?
thanks in advance
***********************************************************************=
********
Forum Note: Use "Reply" to post a response in the discussion forum.
=
I found something about the timestamp problem. IC54415: TIMESTAMP WRAP ASSOCIATED WITH ARCHIVE LEADS TO ARCHIVE RESERVE PAGE DISCREPANCY http://www-01.ibm.com/support/docview.wss?uid=swg1IC54415 is that what you mentioned before?
Dear all
update something new...
I looked the article from IBM and checked the report between onstat -pr and
onstat -g arc. Logical Log Position is the same. Is any idea about thisproblem? it really drive my crazy...
First try: run that ontape -s -L 0 manually and see what happens. Unless
you have redirected the cron script's output you don't know if ontape
reported a problem, was waiting for you to respond to a prompt, or what.
Art
Art S. Kagel
Advanced DataTools (www.advancedatatools.com)
IIUG Board of Directors (art@iiug.org)
See you at the 2010 IIUG Informix Conference
April 25-28, 2010
Overland Park (Kansas City), KS
www.iiug.org/conf
Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Advanced DataTools, the IIUG, nor any other
organization with which I am associated either explicitly, implicitly, or by
inference. Neither do those opinions reflect those of other individuals
affiliated with any entity with which I am affiliated nor those of the
entities themselves.
On Thu, Jan 21, 2010 at 11:27 PM, FAN YANG <fania.yang@gmail.com> wrote:
> Dear all
>
> I got a strange problem about using ontape (on IDS 9.4.FC8 with AIX).
> I wrote a backup script like :
> TAPE=`onstat -c | grep "^TAPEDEV" | awk '{print $2}'`
> echo "Begin ontape backup to $TAPE: `date`"
> ontape -s -L 0 <<!>
> ! echo "Finish ontape backup to $TAPE: `date`"
>
> I use it in the cron job to perform automatic backup and it workes fine all
> the time. Usually it will finish before 6 am. However this morning it is
> still
> runing till 9 am.In the log message, I usually can see the following
> message.
>
> Please mount tape 1 on /IDS9.40/informix/tape and press Return to continue
> ...
> 10 percent done.
> 20 percent done.
> 30 percent done.
> 40 percent done.
> 50 percent done.
> 100 percent done.
>
> Please label this tape as number 1 in the arc tape sequence.
> This tape contains the following logical logs:
>
> 956
>
> Program over
>
> But today I didn't see those information in the log and there is no error
> message in online.log. Besides that I saw that disk had high read/write.
> Because it will make a low performance, I broke the program by manual.
>
> Could anyone can help me to solve this problem?
> thanks in advance
>
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>
--00151747c0aec25279047dc0733f
This is not the problem I was referring to. Please look for very_old_page. John F. Miller III STSM, Support Architect miller3@us.ibm.com 503-578-5645 IBM Informix Dynamic Server (IDS) ids-bounces@iiug.org wrote on 01/21/2010 10:55:12 PM: > [image removed] > > Re: IDS9.4FC8 backup problem [18743] > > FAN YANG > > to: > > ids > > 01/21/2010 10:56 PM > > Sent by: > > ids-bounces@iiug.org > > Please respond to ids > > I found something about the timestamp problem. > IC54415: TIMESTAMP WRAP ASSOCIATED WITH ARCHIVE LEADS TO ARCHIVE RESERVE PAGE > DISCREPANCY > http://www-01.ibm.com/support/docview.wss?uid=swg1IC54415 > > is that what you mentioned before? > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. >
hi, all I tried to put "CCFLAGS" in the ONCONFIG file however the problem is happening again. the log doesn't show any error again. could any one help me that? are we running into another bug issue after I set the CCFLAGS? what can help me to avoid this problem? thanks in advance
On 4 February 2010 07:15, FAN YANG <fania.yang@gmail.com> wrote: > hi, all > > I tried to put "CCFLAGS" in the ONCONFIG file however the problem is happening > again. the log doesn't show any error again. could any one help me that? > are we running into another bug issue after I set the CCFLAGS? > > what can help me to avoid this problem? > > thanks in advance > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > There is no context here! You must includesome details of the original issues as e-mail links to the forum do not thread very well (if at all) due to the forum's habits of appending an index number to the end of the Subject line. Also, I believe it is some time since this issue was raised and discussed. Keith
Thanks for Keith reminded me include all detail about this issue....
Here is some detail of the original issue...
-----------------------------------------------------------------------
hi, all
I tried to put "CCFLAGS" in the ONCONFIG file however the problem is happening
again. the log doesn't show any error again. could any one help me that?
are we running into another bug issue after I set the CCFLAGS?
what can help me to avoid this problem?
thanks in advance,
Fan
-------------------------------------------------------------------------------
This sounds like the very old page issue. There is many post about it,=
but
in short every so often the backup has to update the timestamp of pages=
which
are not accessed for update. This is done by a level 0 archive and can=
general
be viewed by high read and write counts in an onstat -u.
The frequency you will see this depends on you update activity. You co=
uld
see it
today then not for another year. This issue has been removed in new
versions.
John F. Miller III
STSM, Support Architect
miller3@us.ibm.com
503-578-5645
IBM Informix Dynamic Server (IDS)
|------------>
| From: |
|------------>
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|"FAN YANG"
=
|
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|------------>
| To: |
|------------>
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|ids@iiug.org =
=
|
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|------------>
| Date: |
|------------>
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|01/21/2010 08:28 PM =
=
|
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|------------>
| Subject: |
|------------>
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|IDS9.4FC8 backup problem [18741] =
=
|
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|------------>
| Sent by: |
|------------>
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
|ids-bounces@iiug.org =
=
|
>--------------------------------------------------------------------=
-----------------------------------------------------------------------=
-------|
Dear all
I got a strange problem about using ontape (on IDS 9.4.FC8 with AIX).
I wrote a backup script like :
TAPE=3D`onstat -c | grep "^TAPEDEV" | awk '{print $2}'`
echo "Begin ontape backup to $TAPE: `date`"
ontape -s -L 0 <<!
! echo "Finish ontape backup to $TAPE: `date`"
I use it in the cron job to perform automatic backup and it workes fine=
all
the time. Usually it will finish before 6 am. However this morning it i=
s
still
runing till 9 am.In the log message, I usually can see the following
message.
Please mount tape 1 on /IDS9.40/informix/tape and press Return to
continue ...
10 percent done.
20 percent done.
30 percent done.
40 percent done.
50 percent done.
100 percent done.
Please label this tape as number 1 in the arc tape sequence.
This tape contains the following logical logs:
956
Program over
But today I didn't see those information in the log and there is no err=
or
message in online.log. Besides that I saw that disk had high read/write=
.
Because it will make a low performance, I broke the program by manual.
Could anyone can help me to solve this problem?
thanks in advance
On 4 February 2010 08:45, FAN YANG <fania.yang@gmail.com> wrote:
> Thanks for Keith reminded me include all detail about this issue....
> Here is some detail of the original issue...
>
> -----------------------------------------------------------------------
> hi, all
>
> I tried to put "CCFLAGS" in the ONCONFIG file however the problem is
happening
> again. the log doesn't show any error again. could any one help me that?
> are we running into another bug issue after I set the CCFLAGS?
>
> what can help me to avoid this problem?
>
> thanks in advance,
> Fan
>
>
>
-------------------------------------------------------------------------------
>
> This sounds like the very old page issue. There is many post about it,=
> but
> in short every so often the backup has to update the timestamp of pages=
>
> which
> are not accessed for update. This is done by a level 0 archive and can=
>
> general
> be viewed by high read and write counts in an onstat -u.
>
> The frequency you will see this depends on you update activity. You co=
> uld
> see it
> today then not for another year. This issue has been removed in new
> versions.
>
> John F. Miller III
> STSM, Support Architect
> miller3@us.ibm.com
> 503-578-5645
> IBM Informix Dynamic Server (IDS)
> | From: |
> |"FAN YANG"
> | To: |
> |ids@iiug.org =
> | Date: |
> |01/21/2010 08:28 PM =
> | Subject: |
> |IDS9.4FC8 backup problem [18741] =
> | Sent by: |
> |ids-bounces@iiug.org =
> Dear all
>
> I got a strange problem about using ontape (on IDS 9.4.FC8 with AIX).
> I wrote a backup script like :
> TAPE=3D`onstat -c | grep "^TAPEDEV" | awk '{print $2}'`
> echo "Begin ontape backup to $TAPE: `date`"
> ontape -s -L 0 <<!>
> ! echo "Finish ontape backup to $TAPE: `date`"
>
> I use it in the cron job to perform automatic backup and it workes fine=
> all
>
> the time. Usually it will finish before 6 am. However this morning it i=
> s
> still
> runing till 9 am.In the log message, I usually can see the following
> message.
>
> Please mount tape 1 on /IDS9.40/informix/tape and press Return to
> continue ...
> 10 percent done.
> 20 percent done.
> 30 percent done.
> 40 percent done.
> 50 percent done.
> 100 percent done.
>
> Please label this tape as number 1 in the arc tape sequence.
> This tape contains the following logical logs:
>
> 956
>
> Program over
>
> But today I didn't see those information in the log and there is no err=
> or
> message in online.log. Besides that I saw that disk had high read/write=
> ..
> Because it will make a low performance, I broke the program by manual.
>
> Could anyone can help me to solve this problem?
> thanks in advance
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>
Fan
This will not be the 'very old pages' bug. This was only an issue
prior to IDS 7.31 xC8 (where CCFLAGS needed to be set). For later
versions the issue was fixed in the code. I encountered the issue on
my 7.31 engine many years ago but upgraded to 9.4 FC6 (on AIX 5.3)
4years ago, didn't set the parameter, and have never encountered
varying archive times since. I think you need to look elsewhere for a
resolution.
How long does your archive normally take?
What happens if you start an archive manually from the command line?
Has the tape filled up and is the archive waiting for another tape to be
loaded?
Is there a new batch job running on the system when the archive runs.?
Keith
normally it will take 8-9 hours. (we have around 300G data) if I start an archive manually it will show.. Please mount tape 1 on /IDS9.40/informix/tape and press Return to continue ... 10 percent done. .... but sometime it will only show "Please mount tape 1 on /IDS9.40/informix/tape and press Return to continue ..." and no information about what percentage of the job. the tape didn't filled up because it's ok for another day. there is no new batch job running on the system when the archive runs. thanks, Fan >This will not be the 'very old pages' bug. This was only an issue >prior to IDS 7.31 xC8 (where CCFLAGS needed to be set). For later >versions the issue was fixed in the code. I encountered the issue on >my 7.31 engine many years ago but upgraded to 9.4 FC6 (on AIX 5.3) >4years ago, didn't set the parameter, and have never encountered >varying archive times since. I think you need to look elsewhere for a >resolution. >How long does your archive normally take? >What happens if you start an archive manually from the command line? >Has the tape filled up and is the archive waiting for another tape to be loaded? >Is there a new batch job running on the system when the archive runs.? >Keith
On 5 February 2010 03:20, FAN YANG <fania.yang@gmail.com> wrote:
> normally it will take 8-9 hours. (we have around 300G data)
> if I start an archive manually it will show..
> Please mount tape 1 on /IDS9.40/informix/tape and press Return to continue
...
> 10 percent done. ....
>
> but sometime it will only show "Please mount tape 1 on /IDS9.40/informix/tape
> and press Return to continue ..." and no information about what percentage of
> the job.
>
> the tape didn't filled up because it's ok for another day.
>
> there is no new batch job running on the system when the archive runs.
>
> thanks,
> Fan
>
>>This will not be the 'very old pages' bug. This was only an issue
>>prior to IDS 7.31 xC8 (where CCFLAGS needed to be set). For later
>>versions the issue was fixed in the code. I encountered the issue on
>>my 7.31 engine many years ago but upgraded to 9.4 FC6 (on AIX 5.3)
>>4years ago, didn't set the parameter, and have never encountered
>>varying archive times since. I think you need to look elsewhere for a
>>resolution.
>>How long does your archive normally take?
>>What happens if you start an archive manually from the command line?
>>Has the tape filled up and is the archive waiting for another tape to be
> loaded?
>>Is there a new batch job running on the system when the archive runs.?
>
>>Keith
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>
Fan
8-9 hours for 300 Gb seems a long time (I'm archiving 500 Gb in 2 1/2
hours with 70-80 connected OLTP users and batch jobs running).
How is your tape connected?
What are your onconfig tape parameters?
What does 'onstat -g arc' show at intervals during the archive?
keith
On 5 February 2010 03:20, FAN YANG <fania.yang@gmail.com> wrote:
> normally it will take 8-9 hours. (we have around 300G data)
> if I start an archive manually it will show..
> Please mount tape 1 on /IDS9.40/informix/tape and press Return to continue
...
> 10 percent done. ....
>
> but sometime it will only show "Please mount tape 1 on /IDS9.40/informix/tape
> and press Return to continue ..." and no information about what percentage of
> the job.
>
> the tape didn't filled up because it's ok for another day.
>
> there is no new batch job running on the system when the archive runs.
>
> thanks,
> Fan
>
>>This will not be the 'very old pages' bug. This was only an issue
>>prior to IDS 7.31 xC8 (where CCFLAGS needed to be set). For later
>>versions the issue was fixed in the code. I encountered the issue on
>>my 7.31 engine many years ago but upgraded to 9.4 FC6 (on AIX 5.3)
>>4years ago, didn't set the parameter, and have never encountered
>>varying archive times since. I think you need to look elsewhere for a
>>resolution.
>>How long does your archive normally take?
>>What happens if you start an archive manually from the command line?
>>Has the tape filled up and is the archive waiting for another tape to be
> loaded?
>>Is there a new batch job running on the system when the archive runs.?
>
>>Keith
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
> On 5 February 2010 03:28, Keith Simmons wrote:
>8-9 hours for 300 Gb seems a long time (I'm archiving 500 Gb in 2 1/2
>hours with 70-80 connected OLTP users and batch jobs running).
>How is your tape connected?
>What are your onconfig tape parameters?
>What does 'onstat -g arc' show at intervals during the archive?
>keith
hi, we use raw device to put the tape.
and there is some parameter that we set in the onconfig...
TAPEBLK 512
TAPEDEV /IDS9.40/informix/tape
TAPEDEV 307200000if you need more just tell me, thanks
in your previous response, you said "For later
versions the issue was fixed in the code". After checking the technote from
IBM,
it also showes the fix is not enabled by default. so i am not sure if i need
to add the ccflags parameter in the onconfig or not. Now I have already put it
in the onconfig, do i need to disable it?
thanks,
Fan