Logs Full
Hello: I'm working with IDS 7.30 on SCO Open Server.
The LTAPEDEV is set to /dev/null
but I get Logical Log Files are Full -- Backup is Needed.
Then I could not shutdown the Engine.
Restarted the server.
Change the LTAPEDEV to /dev/rmt0
oninit -s
ontape -aBut nothing occurs,
Please help me.
Hi,
what is the setting of the ONCONFIG-parameters LTXHWM and LTXEHWM ???
Can you send the output of 'onstat -l' ???
Are there other hanging ontape processes ???
Regards,
Andreas Kutsche
> -----Ursprüngliche Nachricht-----
> Von: forum.subscriber@iiug.org [mailto:forum.subscriber@iiug.org]Im
> Auftrag von LUIS PANOZZO
> Gesendet: Donnerstag, 03. Februar 2005 18:13
> An: ids@iiug.org
> Betreff: Logs Full [4159]
>
>
> Hello: I'm working with IDS 7.30 on SCO Open Server.
> The LTAPEDEV is set to /dev/null
> but I get Logical Log Files are Full -- Backup is Needed.
>
> Then I could not shutdown the Engine.
> Restarted the server.
> Change the LTAPEDEV to /dev/rmt0
> oninit -s
> ontape -a> But nothing occurs,
>
> Please help me.
>
>
>
Your
onstat -x shows a (already rolling back) transaction beginning with lo=g 1832
=20
Informix Dynamic Server Version 7.30.UC2 -- Fast Recovery -- Up 00:01:35 =
-- 77824 Kbytes
=20
Transactions
address flags userthread locks log begin isolation retrys coordinator=20=
=20=20=20=20=20=20
8646e014 A---- 86448014 0 0 COMMIT 0=20=20=20=20=20=20=20=
=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20
8646e164 A---- 864484c4 0 0 COMMIT 0=20=20=20=20=20=20=20=
=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20
8646e2b4 A---- 86448974 0 0 COMMIT 0=20=20=20=20=20=20=20=
=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20
8646e404 A---- 86448e24 0 0 COMMIT 0=20=20=20=20=20=20=20=
=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20
8646e554 A---- 864492d4 0 0 COMMIT 0=20=20=20=20=20=20=20=
=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20
8646e6a4 A-R-- 864492d4 0 1832 COMMIT 0=20=20=20=20=20=20=20=
=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20
6 active, 128 total, 0 maximum concurrent
=20
=20
Your onstat -l before showed, that the server uses the logfiles with IDs=20
1837=20
....
1843 - current log
1832
....
1836
=20
Therefore your IDS is indefinitely blocked - the server would need a new lo=
gical log to=20
continue with the transaction rollback, but he can't use the next one (1832=
), because
this one has some information about the transaction being rolled back which=
will be=20
needed to perform the rollback.
=20
As far as I know you can only do a restore from a level-0 backup. In theory=
you could
perform a rollforward until the problematic transaction has used a few logf=
iles and let
the rollback complete. But you said your server had LTAPEDEV set to /dev/nu=
ll -=20
therefore you only a level-0 backup.
I hope you have one?
=20
If not - or if you don't want to do the restore - you may call IBM/Informix=
tech support
if they can bring the server Online (perhaps to export the data), but I don=
't know if that's possible.
=20
Maybe you wait for some other advice from the user group, but I don't know =
any workaround.
With newer IDS version (9.40) you could add more logical logs, but not with=
7.30.
=20
Sorry to say another thing, but I think you worsened the situation with swi=
tching logfiles
manually, taking logbackups or stopping/starting the server - your logfil=
es 1836 to 1842=20
are not really filled (only a few percent each). With the standard setting =
of the
LTXHWM and LTXEHWM (40%, 60%) your situation should never occur, if all log=
files
are used to 100% (these parameters define when a transaction rollback start=
s - see=20
documentation).
=20
Regards,
Andreas Kutsche
=20
-----Urspr=FCngliche Nachricht-----
Von: lupasa@sisa.com.bo [mailto:lupasa@sisa.com.bo]
Gesendet: Donnerstag, 03. Februar 2005 20:10
An: KUTSCHE Andreas (HZ422)
Betreff: Re: AW: AW: Logs Full [4159]
Hello:=20
The onstat -x shows me the following:=20
And the engina still in fastrecovery.=20
Thank You.=20
Luis Panozzo Saavedra
Gerente T=E9cnico
lupasa@sisa.com.bo
Soluciones Integrales S. A.
Telf. (591-2)2773378
Fax (591-2)2794538
Edif. Terranova Of. 207 C.Ignacio Cordero 976 San Miguel
www.sisa.com.bo=20
La Paz - Bolivia=20
"Andreas.KUTSCHE@spar.at" <andreas.kutsche@spar.at>=20
03/02/2005 02:38 p.m.=20
Para
<lupasa@sisa.com.bo>=20
cc
Asunto
AW: AW: Logs Full [4159]
=09
Hi,=20
=20=20
your onstat -l shows=20=20=20
=20=20
U-B---- in most lines of the third column, that means all logical logs ar=
e already marked as backed up.=20
=20=20
Does your problem still exist or is your server Online and working again?=
=20
=20=20
If you have still a problem, try onstat -x and examine the forth column=
(log begin) if a large=20
transaction exists that spans all of your logs - in that case I think yo=
ur server is locked=20
and cannot be unlocked (maybe tech support can).=20
=20=20
=20=20
Regards,=20
Andreas=20
=20=20
-----Urspr=FCngliche Nachricht-----
Von: lupasa@sisa.com.bo [mailto:lupasa@sisa.com.bo]
Gesendet: Donnerstag, 03. Februar 2005 19:42
An: KUTSCHE Andreas (HZ422)
Betreff: Re: AW: Logs Full [4159]
Hello:=20
No other ontape process.=20
The onstat -l output is as follows:=20
U-B-...=20=20
Thank You.=20
Luis Panozzo Saavedra
La Paz - Bolivia=20