Cannot drop database in IDS 11.50.FC4
After upgrading from IDS 9.40 to 11.50.FC4 on HP-UX 11.23, the poster's nightly scripts that drop and reload test databases began failing with error 425/107 (lock/in-use) on DROP DATABASE, even though onstat showed no connections or locks. IBM support said the bug was fixed in XC4 and suggested the workaround of running admin("scheduler shutdown") before the drop and restarting it afterwards, which worked but required script changes. One reply suggested permanently disabling the dbScheduler/dbWorker threads by creating $INFORMIXDIR/etc/sysadmin/stop and bouncing the server. No confirmed resolution from the original poster is recorded; the rest of the thread drifted into off-topic bickering.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
We recently upgraded to IDS 11.50.FC4 from 9.40 on HPUX 11.23 and are
experiencing sharp teething pains. We have scripts which run during
off hours to reload test and development databases by shutting down
appservers to drop all connections, then drop the target DB. Then
dbimport/myimport is used to load. Now after all connections have
been dropped we get 425/107 errors when the script attempts to execute
'drop database'. Examining the database when this occurs we find no
connections reported by onstat, and queries vs. system catalog tables
show only empty lock-related tables. Call to IBM tech support
indicates this problem was reported and solved with XC4, with
workaround to shutdown scheduler (execute function admin("scheduler
shutdown")) immediately prior to drop and re-start immediately
afterwards. This works but we'd like to avoid editing our scripts,
and besides, this bug was reported fixed in this version (maybe
platform-specific idiosyncracy?).
We've not been able to reproduce the state that induces the error -- a
database can easily be created immediately and dropped in the
afflicted instance. What else can we check? Anyone understand what's
causing this behavior?
<red_valsen@yahoo.com> wrote in message
news:781c9941-9b42-4475-bfad-bd6c9ec58e51@o14g2000vbo.googlegroups.com...
> We recently upgraded to IDS 11.50.FC4 from 9.40 on HPUX 11.23 and are
> experiencing sharp teething pains. We have scripts which run during
> off hours to reload test and development databases by shutting down
> appservers to drop all connections, then drop the target DB. Then
> dbimport/myimport is used to load. Now after all connections have
> been dropped we get 425/107 errors when the script attempts to execute
> 'drop database'. Examining the database when this occurs we find no
> connections reported by onstat, and queries vs. system catalog tables
> show only empty lock-related tables. Call to IBM tech support
> indicates this problem was reported and solved with XC4, with
> workaround to shutdown scheduler (execute function admin("scheduler
> shutdown")) immediately prior to drop and re-start immediately
> afterwards.
From: "Obnoxio The Clown" <obnoxio@serendipita.com>
Newsgroups: comp.databases.informix
Cc: <informix-list@iiug.org>
Sent: Monday, June 01, 2009 9:39 AM
Subject: Re: {Spam?} RE: Informix Roundup: 30 May
>> Some of the most vocal members of this forum have dragged their heels
>> like motherfuckers to stop people from upgrading to IDS11 till the very
>> last moment, why are you surprised about case studies mentioning IDS 10?
Maybe some of those vocal members are out in the real world, and actually
know what they're doing ...? ;-)
Captain Pedantic wrote: > > Maybe some of those vocal members are out in the real world, and actually > know what they're doing ...? ;-) That's all good and well, but then you can't bitch about a lack of market presence for the latest version on one hand and then strut about your real world smarts on the other. Unless, of course, you're a smug hypocrite. -- Cheers, Obnoxio The Clown http://obotheclown.blogspot.com -- This message has been scanned for viruses and dangerous content by OpenProtect(http://www.openprotect.com), and is believed to be clean.
On Jun 1, 6:53 pm, Obnoxio The Clown <obno...@serendipita.com> wrote: > Captain Pedantic wrote: > > > Maybe some of those vocal members are out in the real world, and actually > > know what they're doing ...? ;-) > > That's all good and well, but then you can't bitch about a lack of > market presence for the latest version on one hand and then strut about > your real world smarts on the other. > > Unless, of course, you're a smug hypocrite. > Pot? Meet Kettle. ;-) Sorry OTC looks like you put your foot in your mouth again. Two different topics like apples and oranges. Its not the version of the win, but the 'freshness' of the win. Scott P brought us a very cool and timely win a couple of years ago, but IBM missed the opportunity. Were they to try and use it to show how green IDS is, it would be pretty much worthless.
Ian Michael Gumby wrote: > On Jun 1, 6:53 pm, Obnoxio The Clown <obno...@serendipita.com> wrote: >> Captain Pedantic wrote: >> >>> Maybe some of those vocal members are out in the real world, and actually >>> know what they're doing ...? ;-) >> That's all good and well, but then you can't bitch about a lack of >> market presence for the latest version on one hand and then strut about >> your real world smarts on the other. >> >> Unless, of course, you're a smug hypocrite. >> > Pot? Meet Kettle. ;-) > > Sorry OTC looks like you put your foot in your mouth again. > Two different topics like apples and oranges. > > Its not the version of the win, but the 'freshness' of the win. That's not what you said before. Before you said "there are no wins". Now you're saying "the wins aren't fresh enough." So if I find you some fresh wins, then you'll say "it's not enough in quantity" or find some other thing. It's like arguing with a woman: you know you're right, but also know "she" is going to cheat until you give up. -- Cheers, Obnoxio The Clown http://obotheclown.blogspot.com -- This message has been scanned for viruses and dangerous content by OpenProtect(http://www.openprotect.com), and is believed to be clean.
red_valsen@yahoo.com schrieb:
> We recently upgraded to IDS 11.50.FC4 from 9.40 on HPUX 11.23 and are
> experiencing sharp teething pains. We have scripts which run during
> off hours to reload test and development databases by shutting down
> appservers to drop all connections, then drop the target DB. Then
> dbimport/myimport is used to load. Now after all connections have
> been dropped we get 425/107 errors when the script attempts to execute
> 'drop database'. Examining the database when this occurs we find no
> connections reported by onstat, and queries vs. system catalog tables
> show only empty lock-related tables. Call to IBM tech support
> indicates this problem was reported and solved with XC4, with
> workaround to shutdown scheduler (execute function admin("scheduler
> shutdown")) immediately prior to drop and re-start immediately
> afterwards. This works but we'd like to avoid editing our scripts,
> and besides, this bug was reported fixed in this version (maybe
> platform-specific idiosyncracy?).
>
> We've not been able to reproduce the state that induces the error -- a
> database can easily be created immediately and dropped in the
> afflicted instance. What else can we check? Anyone understand what's
> causing this behavior?
Hi,
if 'scheduler' is the new dbScheduler, then I have this in the output
of onstat -m
03:22:19 On-Line Mode
03:22:20 SCHAPI: "INFORMIXDIR/etc/sysadmin/stop" file is present. Bypassing dbScheduler and dbWorker threads startup.
As I did not understand what the sysadmin database is good for,
the first thing for me to do was to deactivate it until I test what
happens when I gonna drop it altogether.
Some searching brought how to avoid having dbScheduler and
dbWorker active:
create a file with the name 'stop'
touch $INFORMIXDIR/etc/sysadmin/stop
Then bounce the server
onmode -ky && oninit -v
Maybe worth a try...
dic_k
P.S.: remarks: Yes, I did read SQL syntax V11 docs, pg 6-5 ff.
and no, I am not on MS Windows or Vista, and if so I'd have
cygwin running including X Windows (to use onperf) and to
have cron and at.
--
Richard Kofler
SOLID STATE EDV
Dienstleistungen GmbH
Vienna/Austria/Europe