IDS 7.30 long transaction problem
A 7.30 instance (SAP) was wedged in LONGTX/CKPT REQ with all logical logs full and no backups. Advice: back up the logical logs (onbar -b -l, or ontape -a with LTAPEDEV=/dev/null) so the long-transaction rollback can finish, then raise log space, lower LTXHWM/LTXEHWM and commit more often. The poster found log backups still aborted even after killing processes and restarting. Others explained the logs can't be backed up until the rollback completes; if LBU_PRESERVE=1 you can free space, otherwise Informix tech support must truncate the last log. No confirmation of recovery is recorded.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
Hi,
I actually have a database frozen in a long transaction mode.
The fact is don't have any backup to restore it.
I know why the database is in this state. What I need is to know what to
do to keep it alive.
Here are results of some onstat things. onmode -c and onmode-c unblock
don't do anything.
An onbar doesn't do anything either.
What will happen if I kill every process and restart the base ? Maybe I
need to do something while the database is still up and running ?
Thanks,
informix> onstat -
Informix Dynamic Server Version 7.30.UC7XB -- On-Line (CKPT REQ)
(LONGTX) -- Up 43 days 02:19:45 -- 617296 Kbytes
Blocked:CKPT LONGTX
informix> onstat -l
Informix Dynamic Server Version 7.30.UC7XB -- On-Line (CKPT REQ)
(LONGTX) -- Up 43 days 02:20:00 -- 617296 Kbytes
Blocked:CKPT LONGTX
Physical Logging
Buffer bufused bufsize numpages numwrits pages/io
P-2 0 512 657177 2814 233.54
phybegin physize phypos phyused %used
1500035 100000 11706 46445 46.45
Logical Logging
Buffer bufused bufsize numrecs numpages numwrits recs/pages pages/io
L-2 14 16 8973198 841536 300554 10.7 2.8
Subsystem numrecs Log Space used
OLDRSAM 8973198 1301223528
Buffer Waiting
Buffer ioproc flags
L-2 2124af20 0x1 0
address number flags uniqid begin size used %used
11131838 1 U------ 2254 200035 25000 25000 100.00
11131854 2 U-----L 2255 2061dd 25000 25000 100.00
11131870 3 U---C-- 2256 214c3d 25000 25000 100.00
1113188c 4 U-B---- 2251 21ade5 25000 25000 100.00
111318a8 5 U-B---- 2252 220f8d 25000 25000 100.00
111318c4 6 U------ 2253 20c385 25000 25000 100.00
informix> onstat -m
Informix Dynamic Server Version 7.30.UC7XB -- On-Line (CKPT REQ)
(LONGTX) -- Up 43 days 02:20:29 -- 617296 Kbytes
Blocked:CKPT LONGTX
Message Log File: /informix/BWD/online.suriname.bwd.log
07:09:06 Aborting Long Transaction: tx 0x1b1715e8 username: sapr3 uid:2015
07:10:14 Lock table overflow - user id 2015, session id 23263
07:10:14 Lock table overflow - user id 2015, session id 27023
07:10:17 Lock table overflow - user id 2015, session id 29407
07:10:17 Lock table overflow - user id 2015, session id 29407
07:10:23 Lock table overflow - user id 2015, session id 27023
07:10:52 Lock table overflow - user id 2010, session id 41064
07:10:52 Logical Log 2253 - Backup Aborted
Aborted by client.
07:10:57 Lock table overflow - user id 2010, session id 41064
07:10:57 Process exited with return code 148: /bin/sh /bin/sh -c/informix/BWD/etc/log_full.sh 2 23 "Logical Log 2253 Complete." "Logical
Log 2253 Complete."
07:11:14 Lock table overflow - user id 2041, session id 41071
07:11:15 Lock table overflow - user id 2041, session id 41072
07:11:17 Lock table overflow - user id 2015, session id 29407
07:11:18 Lock table overflow - user id 2015, session id 47
07:11:26 Logical Log 2255 Complete.
07:11:27 Lock table overflow - user id 2010, session id 41074
07:20:43 Logical Log Files are Full -- Backup is Needed
--
_______________________________________________
Sebastien THOMAS none networks
System Engineer freesbee
geo:153, rue Saint-Denis, 75002 Paris, France
vox:+33 1 45 08 23 10 - fax:+33 1 45 08 25 29
mailto:sebastien.thomas@none.net
You need to back up your logs so that Informix can complete the rollback of
the "long" transaction - until then, your instance will be "locked up".
Use "onbar -b -l" if you have configured onbar to run. If not, use ontape
to backup the logs. If you don't care about your logs, set LTAPEDEV to
/dev/null in your ONCONFIG file and run "ontape -a". Once logs are backed
up, Informix will continue with the rollback (check for activity by
initializing stats "onstat -z" and running onstat -p).
You need to avoid situations like this in the future using either or both
of the following :
1. Increase the total size of your logs (number of logs x logsize)
2. Get your developers to reduce the size of their transactions by
COMMITting more often, if possible.
All the best
Rudy
Sebastien THOMAS wrote:
> Hi,
>
> I actually have a database frozen in a long transaction mode.
> The fact is don't have any backup to restore it.
> I know why the database is in this state. What I need is to know what to
> do to keep it alive.
> Here are results of some onstat things. onmode -c and onmode-c unblock
> don't do anything.
> An onbar doesn't do anything either.
>
> What will happen if I kill every process and restart the base ? Maybe I
> need to do something while the database is still up and running ?
>
After you solve the problem as Rudy's instruction, you have better to
reduce the value of "LTXHWM" and "LTXEHWM" what parameters in onconfig file.
Otherwise it will cause system crash if there is not enough log space for
roll back long transaction successful.
Eric Chen
Trainer
Education Service Dept.
Informix Software Taiwan Inc.
"Sebastien THOMAS" <Sebastien.THOMAS@none.net> ?????
news:3A018F96.DC2BBAC1@none.net...
> Hi,
>
> I actually have a database frozen in a long transaction mode.
> The fact is don't have any backup to restore it.
> I know why the database is in this state. What I need is to know what to
> do to keep it alive.
> Here are results of some onstat things. onmode -c and onmode-c unblock
> don't do anything.
> An onbar doesn't do anything either.
>
> What will happen if I kill every process and restart the base ? Maybe I
> need to do something while the database is still up and running ?
>
> Thanks,
>
>
>
> informix> onstat -
>
> Informix Dynamic Server Version 7.30.UC7XB -- On-Line (CKPT REQ)
> (LONGTX) -- Up 43 days 02:19:45 -- 617296 Kbytes
> Blocked:CKPT LONGTX
>
>
> informix> onstat -l
>
> Informix Dynamic Server Version 7.30.UC7XB -- On-Line (CKPT REQ)
> (LONGTX) -- Up 43 days 02:20:00 -- 617296 Kbytes
> Blocked:CKPT LONGTX
>
> Physical Logging
> Buffer bufused bufsize numpages numwrits pages/io
> P-2 0 512 657177 2814 233.54
> phybegin physize phypos phyused %used
> 1500035 100000 11706 46445 46.45
>
> Logical Logging
> Buffer bufused bufsize numrecs numpages numwrits recs/pages pages/io
> L-2 14 16 8973198 841536 300554 10.7 2.8
> Subsystem numrecs Log Space used
> OLDRSAM 8973198 1301223528
>
> Buffer Waiting
> Buffer ioproc flags
> L-2 2124af20 0x1 0
>
> address number flags uniqid begin size used %used
> 11131838 1 U------ 2254 200035 25000 25000 100.00
> 11131854 2 U-----L 2255 2061dd 25000 25000 100.00
> 11131870 3 U---C-- 2256 214c3d 25000 25000 100.00
> 1113188c 4 U-B---- 2251 21ade5 25000 25000 100.00
> 111318a8 5 U-B---- 2252 220f8d 25000 25000 100.00
> 111318c4 6 U------ 2253 20c385 25000 25000 100.00
>
>
>
> informix> onstat -m
>
> Informix Dynamic Server Version 7.30.UC7XB -- On-Line (CKPT REQ)
> (LONGTX) -- Up 43 days 02:20:29 -- 617296 Kbytes
> Blocked:CKPT LONGTX
>
> Message Log File: /informix/BWD/online.suriname.bwd.log
> 07:09:06 Aborting Long Transaction: tx 0x1b1715e8 username: sapr3 uid:> 2015
> 07:10:14 Lock table overflow - user id 2015, session id 23263
> 07:10:14 Lock table overflow - user id 2015, session id 27023
> 07:10:17 Lock table overflow - user id 2015, session id 29407
> 07:10:17 Lock table overflow - user id 2015, session id 29407
> 07:10:23 Lock table overflow - user id 2015, session id 27023
> 07:10:52 Lock table overflow - user id 2010, session id 41064
> 07:10:52 Logical Log 2253 - Backup Aborted
> Aborted by client.
> 07:10:57 Lock table overflow - user id 2010, session id 41064
> 07:10:57 Process exited with return code 148: /bin/sh /bin/sh -c> /informix/BWD/etc/log_full.sh 2 23 "Logical Log 2253 Complete." "Logical
> Log 2253 Complete."
> 07:11:14 Lock table overflow - user id 2041, session id 41071
> 07:11:15 Lock table overflow - user id 2041, session id 41072
> 07:11:17 Lock table overflow - user id 2015, session id 29407
> 07:11:18 Lock table overflow - user id 2015, session id 47
> 07:11:26 Logical Log 2255 Complete.
> 07:11:27 Lock table overflow - user id 2010, session id 41074
> 07:20:43 Logical Log Files are Full -- Backup is Needed>
>
>
>
>
>
>
> --
> _______________________________________________
> Sebastien THOMAS none networks
> System Engineer freesbee
>
> geo:153, rue Saint-Denis, 75002 Paris, France
> vox:+33 1 45 08 23 10 - fax:+33 1 45 08 25 29
> mailto:sebastien.thomas@none.net
>
>
>
Hi,
I tried something like that... I can't backup anything.
I finaly tried a onmode -ky, to shut down the database. That was not working
at all. I then killed every process, and restarted the database,
It reloads everything, as you can see in the logs :
Fri Nov 3 14:50:15 2000
14:50:15 Event alarms enabled. ALARMPROG = '/informix/BWD/etc/log_full.sh'
14:50:20 DR: DRAUTO is 0 (Off)
14:50:20 Informix Dynamic Server Version 7.30.UC7XB Software Serial NumberACN#J266875
14:50:21 WARNING! LTXHWM (70) and/or LTXEHWM (80) config parameters exceed recommended values of 50 and 60. This may
cause On-Line to wait too long before rolling
back a long transaction and allow it to completely
fill the logs. See the Dynamic Server Administrators Guide
and $INFORMIXDIR/release/SERVERS_7.1 for more
information on resetting LTXHWM and LTXEHWM.
14:50:21 Informix Dynamic Server Initialized -- Shared Memory Initialized.
14:50:21 Physical Recovery Started.
14:50:31 Physical Recovery Complete: 46445 Pages Restored.
14:50:31 Logical Recovery Started.
14:57:03 Logical Log Files are Full -- Backup is Needed
15:04:20 Logical Log 2253 - Backup Started
15:04:47 Logical Log 2253 - Backup Aborted
Aborted by client.
15:08:52 Logical Log 2253 - Backup Started
15:18:58 Requested shared memory segment size rounded from 337700KB to337712KB
Fri Nov 3 15:18:59 2000
15:18:59 Event alarms enabled. ALARMPROG = '/informix/BWD/etc/log_full.sh'
15:19:04 DR: DRAUTO is 0 (Off)
15:19:04 Informix Dynamic Server Version 7.30.UC7XB Software Serial NumberACN#J266875
15:19:04 WARNING! LTXHWM (70) and/or LTXEHWM (80) config parameters exceed recommended values of 50 and 60. This may
cause On-Line to wait too long before rolling
back a long transaction and allow it to completely
fill the logs. See the Dynamic Server Administrators Guide
and $INFORMIXDIR/release/SERVERS_7.1 for more
information on resetting LTXHWM and LTXEHWM.
15:19:05 Informix Dynamic Server Initialized -- Shared Memory Initialized.
15:19:05 Physical Recovery Started.
15:19:15 Physical Recovery Complete: 46445 Pages Restored.
15:19:15 Logical Recovery Started.
15:26:03 Logical Log Files are Full -- Backup is Needed
I tried an onbar, and I also changed LTAPEDEV to dev/null and tried an ontape.
None of this work. May the database so full that it can't even backup ?? :)
What if I add some logical logs, and do a fake backup ??
I don't care of things in the logs. I just want to have my database alive.
I'll check for data concistancy later.
For information. I'm not in charge of this database. This database is used for
SAP, and our SAP guy's configured it this way. Now it's fucked up, and I have
to take them out of the shit they jumped.
Thanks for your precious help.
Rudy Fernandes wrote:
> You need to back up your logs so that Informix can complete the rollback of
> the "long" transaction - until then, your instance will be "locked up".
> Use "onbar -b -l" if you have configured onbar to run. If not, use ontape
> to backup the logs. If you don't care about your logs, set LTAPEDEV to
> /dev/null in your ONCONFIG file and run "ontape -a". Once logs are backed
> up, Informix will continue with the rollback (check for activity by
> initializing stats "onstat -z" and running onstat -p).
>
> You need to avoid situations like this in the future using either or both
> of the following :
> 1. Increase the total size of your logs (number of logs x logsize)
> 2. Get your developers to reduce the size of their transactions by
> COMMITting more often, if possible.
>
> All the best
> Rudy
>
> Sebastien THOMAS wrote:
>
> > Hi,
> >
> > I actually have a database frozen in a long transaction mode.
> > The fact is don't have any backup to restore it.
> > I know why the database is in this state. What I need is to know what to
> > do to keep it alive.
> > Here are results of some onstat things. onmode -c and onmode-c unblock
> > don't do anything.
> > An onbar doesn't do anything either.
> >
> > What will happen if I kill every process and restart the base ? Maybe I
> > need to do something while the database is still up and running ?
> >
--
_______________________________________________
Sebastien THOMAS none networks
System Engineer freesbee
geo:153, rue Saint-Denis, 75002 Paris, France
vox:+33 1 45 08 23 10 - fax:+33 1 45 08 25 29
mailto:sebastien.thomas@none.net
In article <3A01AD3C.662BF6ED@americasm01.nt.com>,
Rudy Fernandes <rferdy@americasm01.nt.com> wrote:
>
> You need to back up your logs so that Informix can complete the
rollback of
> the "long" transaction - until then, your instance will be "locked
up".
> Use "onbar -b -l" if you have configured onbar to run. If not, use
ontape> to backup the logs. If you don't care about your logs, set LTAPEDEV to
> /dev/null in your ONCONFIG file and run "ontape -a". Once logs are
backed
> up, Informix will continue with the rollback (check for activity by
> initializing stats "onstat -z" and running onstat -p).
>
> You need to avoid situations like this in the future using either or
both
> of the following :
> 1. Increase the total size of your logs (number of logs x logsize)
> 2. Get your developers to reduce the size of their transactions by
> COMMITting more often, if possible.
>
> All the best
> Rudy
>
> Sebastien THOMAS wrote:
>
> > Hi,
> >
> > I actually have a database frozen in a long transaction mode.
> > The fact is don't have any backup to restore it.
> > I know why the database is in this state. What I need is to know
what to
> > do to keep it alive.
> > Here are results of some onstat things. onmode -c and onmode-c
unblock
> > don't do anything.
> > An onbar doesn't do anything either.
> >
> > What will happen if I kill every process and restart the base ?
Maybe I
> > need to do something while the database is still up and running ?
> >
>
>
We know the following.
1. The logs are full.
2. The system is trying to rollback a long transaction.
3. His LTX watermarks are quite high.
I think the next logical log contains part of the transaction being
rolled back. IIRC you can not backup the logs until all the
transactions in it are completed(the rolls back needs to complete for
that), and you cant finish rolling back until you have logs to backup
to.
If LBU_PRESERVE is set to 1, he can add logs. Otherwise the only thing
I can think of to do is call tech support and see if they have any way
to add logs or something of that sort.
Hope this helps,
Will
Sent via Deja.com http://www.deja.com/
Before you buy.
Oh bother.
So from now on you'll be doing L0 archives and continuous logging then
.......
Malc
"Sebastien THOMAS" <Sebastien.THOMAS@none.net> wrote in message
news:3A018F96.DC2BBAC1@none.net...
> Hi,
>
> I actually have a database frozen in a long transaction mode.
> The fact is don't have any backup to restore it.
> I know why the database is in this state. What I need is to know what to
> do to keep it alive.
> Here are results of some onstat things. onmode -c and onmode-c unblock
> don't do anything.
> An onbar doesn't do anything either.
>
> What will happen if I kill every process and restart the base ? Maybe I
> need to do something while the database is still up and running ?
>
> Thanks,
>
>
>
> informix> onstat -
>
> Informix Dynamic Server Version 7.30.UC7XB -- On-Line (CKPT REQ)
> (LONGTX) -- Up 43 days 02:19:45 -- 617296 Kbytes
> Blocked:CKPT LONGTX
>
>
> informix> onstat -l
>
> Informix Dynamic Server Version 7.30.UC7XB -- On-Line (CKPT REQ)
> (LONGTX) -- Up 43 days 02:20:00 -- 617296 Kbytes
> Blocked:CKPT LONGTX
>
> Physical Logging
> Buffer bufused bufsize numpages numwrits pages/io
> P-2 0 512 657177 2814 233.54
> phybegin physize phypos phyused %used
> 1500035 100000 11706 46445 46.45
>
> Logical Logging
> Buffer bufused bufsize numrecs numpages numwrits recs/pages pages/io
> L-2 14 16 8973198 841536 300554 10.7 2.8
> Subsystem numrecs Log Space used
> OLDRSAM 8973198 1301223528
>
> Buffer Waiting
> Buffer ioproc flags
> L-2 2124af20 0x1 0
>
> address number flags uniqid begin size used %used
> 11131838 1 U------ 2254 200035 25000 25000 100.00
> 11131854 2 U-----L 2255 2061dd 25000 25000 100.00
> 11131870 3 U---C-- 2256 214c3d 25000 25000 100.00
> 1113188c 4 U-B---- 2251 21ade5 25000 25000 100.00
> 111318a8 5 U-B---- 2252 220f8d 25000 25000 100.00
> 111318c4 6 U------ 2253 20c385 25000 25000 100.00
>
>
>
> informix> onstat -m
>
> Informix Dynamic Server Version 7.30.UC7XB -- On-Line (CKPT REQ)
> (LONGTX) -- Up 43 days 02:20:29 -- 617296 Kbytes
> Blocked:CKPT LONGTX
>
> Message Log File: /informix/BWD/online.suriname.bwd.log
> 07:09:06 Aborting Long Transaction: tx 0x1b1715e8 username: sapr3 uid:> 2015
> 07:10:14 Lock table overflow - user id 2015, session id 23263
> 07:10:14 Lock table overflow - user id 2015, session id 27023
> 07:10:17 Lock table overflow - user id 2015, session id 29407
> 07:10:17 Lock table overflow - user id 2015, session id 29407
> 07:10:23 Lock table overflow - user id 2015, session id 27023
> 07:10:52 Lock table overflow - user id 2010, session id 41064
> 07:10:52 Logical Log 2253 - Backup Aborted
> Aborted by client.
> 07:10:57 Lock table overflow - user id 2010, session id 41064
> 07:10:57 Process exited with return code 148: /bin/sh /bin/sh -c> /informix/BWD/etc/log_full.sh 2 23 "Logical Log 2253 Complete." "Logical
> Log 2253 Complete."
> 07:11:14 Lock table overflow - user id 2041, session id 41071
> 07:11:15 Lock table overflow - user id 2041, session id 41072
> 07:11:17 Lock table overflow - user id 2015, session id 29407
> 07:11:18 Lock table overflow - user id 2015, session id 47
> 07:11:26 Logical Log 2255 Complete.
> 07:11:27 Lock table overflow - user id 2010, session id 41074
> 07:20:43 Logical Log Files are Full -- Backup is Needed>
>
>
>
>
>
>
> --
> _______________________________________________
> Sebastien THOMAS none networks
> System Engineer freesbee
>
> geo:153, rue Saint-Denis, 75002 Paris, France
> vox:+33 1 45 08 23 10 - fax:+33 1 45 08 25 29
> mailto:sebastien.thomas@none.net
>
>
>
The real problem seems to be that the logical log archives are failing.
This could be because you do not have enough logical log space left to note
the archive. If you were smart enough to set LBU_PRESERVE to 1, shutdown
the engine, reset it to zero, restart and backup the logs, don't forget to
shutdown again and reset LBU_PRESERVE to 1. If it is already set to zero
the you are partially hosed. Call Informix tech support and sign a dial-up
agreement. Then they can login and truncate your last logical log so you
can recover at least to the beginning of that log.
Either way do the following in the future:
set LBU_PRESERVE 1
increase the number of logical logs
reduce LTXHWM and LTXEHWM a bit.
Art S. Kagel
Sebastien THOMAS wrote:
>
> Hi,
>
> I tried something like that... I can't backup anything.
> I finaly tried a onmode -ky, to shut down the database. That was not working
> at all. I then killed every process, and restarted the database,
> It reloads everything, as you can see in the logs :
>
> Fri Nov 3 14:50:15 2000
>
> 14:50:15 Event alarms enabled. ALARMPROG = '/informix/BWD/etc/log_full.sh'
> 14:50:20 DR: DRAUTO is 0 (Off)
> 14:50:20 Informix Dynamic Server Version 7.30.UC7XB Software Serial Number> ACN#J266875
> 14:50:21 WARNING! LTXHWM (70) and/or LTXEHWM (80) config parameters> exceed recommended values of 50 and 60. This may
> cause On-Line to wait too long before rolling
> back a long transaction and allow it to completely
> fill the logs. See the Dynamic Server Administrators Guide
> and $INFORMIXDIR/release/SERVERS_7.1 for more
> information on resetting LTXHWM and LTXEHWM.
>
> 14:50:21 Informix Dynamic Server Initialized -- Shared Memory Initialized.
> 14:50:21 Physical Recovery Started.
> 14:50:31 Physical Recovery Complete: 46445 Pages Restored.
> 14:50:31 Logical Recovery Started.
> 14:57:03 Logical Log Files are Full -- Backup is Needed
>
> 15:04:20 Logical Log 2253 - Backup Started
> 15:04:47 Logical Log 2253 - Backup Aborted
> Aborted by client.
> 15:08:52 Logical Log 2253 - Backup Started
> 15:18:58 Requested shared memory segment size rounded from 337700KB to> 337712KB
>
> Fri Nov 3 15:18:59 2000
>
> 15:18:59 Event alarms enabled. ALARMPROG = '/informix/BWD/etc/log_full.sh'
> 15:19:04 DR: DRAUTO is 0 (Off)
> 15:19:04 Informix Dynamic Server Version 7.30.UC7XB Software Serial Number> ACN#J266875
> 15:19:04 WARNING! LTXHWM (70) and/or LTXEHWM (80) config parameters> exceed recommended values of 50 and 60. This may
> cause On-Line to wait too long before rolling
> back a long transaction and allow it to completely
> fill the logs. See the Dynamic Server Administrators Guide
> and $INFORMIXDIR/release/SERVERS_7.1 for more
> information on resetting LTXHWM and LTXEHWM.
>
> 15:19:05 Informix Dynamic Server Initialized -- Shared Memory Initialized.
> 15:19:05 Physical Recovery Started.
> 15:19:15 Physical Recovery Complete: 46445 Pages Restored.
> 15:19:15 Logical Recovery Started.
> 15:26:03 Logical Log Files are Full -- Backup is Needed>
> I tried an onbar, and I also changed LTAPEDEV to dev/null and tried an ontape.
> None of this work. May the database so full that it can't even backup ?? :)
> What if I add some logical logs, and do a fake backup ??
>
> I don't care of things in the logs. I just want to have my database alive.
> I'll check for data concistancy later.
> For information. I'm not in charge of this database. This database is used for
> SAP, and our SAP guy's configured it this way. Now it's fucked up, and I have
> to take them out of the shit they jumped.
>
> Thanks for your precious help.
>
> Rudy Fernandes wrote:
>
> > You need to back up your logs so that Informix can complete the rollback of
> > the "long" transaction - until then, your instance will be "locked up".
> > Use "onbar -b -l" if you have configured onbar to run. If not, use ontape
> > to backup the logs. If you don't care about your logs, set LTAPEDEV to
> > /dev/null in your ONCONFIG file and run "ontape -a". Once logs are backed
> > up, Informix will continue with the rollback (check for activity by
> > initializing stats "onstat -z" and running onstat -p).
> >
> > You need to avoid situations like this in the future using either or both
> > of the following :
> > 1. Increase the total size of your logs (number of logs x logsize)
> > 2. Get your developers to reduce the size of their transactions by
> > COMMITting more often, if possible.
> >
> > All the best
> > Rudy
> >
> > Sebastien THOMAS wrote:
> >
> > > Hi,
> > >
> > > I actually have a database frozen in a long transaction mode.
> > > The fact is don't have any backup to restore it.
> > > I know why the database is in this state. What I need is to know what to
> > > do to keep it alive.
> > > Here are results of some onstat things. onmode -c and onmode-c unblock
> > > don't do anything.
> > > An onbar doesn't do anything either.
> > >
> > > What will happen if I kill every process and restart the base ? Maybe I
> > > need to do something while the database is still up and running ?
> > >
>
> --
> _______________________________________________
> Sebastien THOMAS none networks
> System Engineer freesbee
>
> geo:153, rue Saint-Denis, 75002 Paris, France
> vox:+33 1 45 08 23 10 - fax:+33 1 45 08 25 29
> mailto:sebastien.thomas@none.net