IDS 9.21
Windows 2000 Advanced Server sp2

Brief history of whine
Use Informix as backend to BaaN ERP system.
Backend to BaaN ERP

Restored live onto test server (approx 216 Gb)

Did a bit of work, including testing dropping some data in reporting tables
as that's now been moved to (SQLServer) datawarehouse (Oh, the flames, the
flames) which reduced the amount of used space quite significantly.

Then did some buggering about with the blocksize on the backup (changed from
16 to 1024 in stages to see if it noticibly improved backup times. Seemed
to), looked at reducing fragmentation in small table trivial test kind of
way.

Backup is to a disc file on the same server. Live and test environments are
an almost exact copy of each other (other then BaaN configuration issues).

Obviously, next stage is to test the restore, which I'd assumed would be a
no brainer. Turned out it was, but not in the way I'd expected.
While the backup looked sweet, finished OK - no errors or issues I can see -
it gets to a point in the restore when it asks for a tape 2.

Now - given that the restore is from a backup to a single disc file, and
that's from a backup to the same single disc file - I find that surprising.
Obviously - there is no tape 2. I *could* <Ctrl - C> but as fate would have
it we've had expensive consultants doing expensive stuff - and
communications breakdown is an unpopular phrase with senior management
expecting a demo.

Question is - is it hosed - or is there a retrieval mechanism anyone here
can give me? All I know is that any time I've stopped a restore
inappropriately (it's happened - we do about 4-6 test DR's a year) - it's
back to square one. Suspect that square one this time's going to bring me
to the same problem point.

Anyway, any advice welcome, if it saves my job you may even get an invite to
my wedding. If you put it on e-bay you die.

On Wed, 10 Nov 2004 23:45:44 GMT, "Andrew" <andrew.gothard1@<nospam
ta}ntlworld.com> wrote:

>IDS 9.21
>Windows 2000 Advanced Server sp2
>
>Brief history of whine
>Use Informix as backend to BaaN ERP system.
>Backend to BaaN ERP
>
>Restored live onto test server (approx 216 Gb)
>

... snipped ..

>Backup is to a disc file on the same server. Live and test environments are
>an almost exact copy of each other (other then BaaN configuration issues).
>
>Obviously, next stage is to test the restore, which I'd assumed would be a
>no brainer. Turned out it was, but not in the way I'd expected.
>While the backup looked sweet, finished OK - no errors or issues I can see -
>it gets to a point in the restore when it asks for a tape 2.
>

Was TAPESIZE changed as well?

>Now - given that the restore is from a backup to a single disc file, and
>that's from a backup to the same single disc file - I find that surprising.
>Obviously - there is no tape 2. I *could* <Ctrl - C> but as fate would have
>it we've had expensive consultants doing expensive stuff - and
>communications breakdown is an unpopular phrase with senior management
>expecting a demo.
>

How big is the disk file backup? I don't think that 9.21 allowed for
disk sizes greater than 2G, but this is Windows here, so . . . . ?

>Question is - is it hosed - or is there a retrieval mechanism anyone here
>can give me? All I know is that any time I've stopped a restore
>inappropriately (it's happened - we do about 4-6 test DR's a year) - it's
>back to square one. Suspect that square one this time's going to bring me
>to the same problem point.
>

JWC

"Andrew" <andrew.gothard1@<nospam ta}ntlworld.com> wrote in message
news:IKxkd.807$fI4.191@newsfe3-gui.ntli.net...
> IDS 9.21
> Windows 2000 Advanced Server sp2
> Now - given that the restore is from a backup to a single disc file, and
> that's from a backup to the same single disc file - I find that
> surprising.
> Obviously - there is no tape 2.

How big is the archive to tape that you *do* have? And what is TAPESIZE set
to in the %ONCONFIG% file?
I'm thinking here that you culd in fact, unknowingly, have written the 2nd
part of the archive to the same disk area, although you'd have had to be
quite drunk to do that inadvertently (and on UNIX at least the database has
rudimentary checks to prevent this).

I *could* <Ctrl - C> but as fate would have
> it we've had expensive consultants doing expensive stuff

You haven't had the right expensive consultants in. We can out-expensive
most people ... ;-)

--
Neil Truby t:01932 724027
Director m:07798 811708
Ardenta Limited e:neil.truby@ardenta.com

Andrew <andrew.gothard1@ wrote:
> IDS 9.21
> Windows 2000 Advanced Server sp2
>
> Brief history of whine
> Use Informix as backend to BaaN ERP system.
> Backend to BaaN ERP
>
> Restored live onto test server (approx 216 Gb)
>
> Did a bit of work, including testing dropping some data in reporting tables
> as that's now been moved to (SQLServer) datawarehouse (Oh, the flames, the
> flames) which reduced the amount of used space quite significantly.
>
> Then did some buggering about with the blocksize on the backup (changed from
> 16 to 1024 in stages to see if it noticibly improved backup times. Seemed
> to), looked at reducing fragmentation in small table trivial test kind of
> way.
>
> Backup is to a disc file on the same server. Live and test environments are
> an almost exact copy of each other (other then BaaN configuration issues).
>
> Obviously, next stage is to test the restore, which I'd assumed would be a
> no brainer. Turned out it was, but not in the way I'd expected.
> While the backup looked sweet, finished OK - no errors or issues I can see -
> it gets to a point in the restore when it asks for a tape 2.

There are a few reason for the ontape process to ask for another
tape. ontape encountered an EOF while reading, TAPESIZE was
encountered (even if using a disk file).

1. The first thing to do is to check the disk file size against the
TAPESIZE paramater in your onconfig.

2. Validate the archive using the archecker utility. If this utility
indicates that all data is there then it can be restored even if
ontape is currently not restoring the data correctly.

Hope this helps,
John

>
> Now - given that the restore is from a backup to a single disc file, and
> that's from a backup to the same single disc file - I find that surprising.
> Obviously - there is no tape 2. I *could* <Ctrl - C> but as fate would have
> it we've had expensive consultants doing expensive stuff - and
> communications breakdown is an unpopular phrase with senior management
> expecting a demo.
>
> Question is - is it hosed - or is there a retrieval mechanism anyone here
> can give me? All I know is that any time I've stopped a restore
> inappropriately (it's happened - we do about 4-6 test DR's a year) - it's
> back to square one. Suspect that square one this time's going to bring me
> to the same problem point.
>
> Anyway, any advice welcome, if it saves my job you may even get an invite to
> my wedding. If you put it on e-bay you die.
>
>

>
> Was TAPESIZE changed as well?

No

>
> How big is the disk file backup? I don't think that 9.21 allowed for
> disk sizes greater than 2G, but this is Windows here, so . . . . ?
>
216 Gb - allowed on Windows

"Andrew" <andrew.gothard1@<nospam ta}ntlworld.com> wrote in message
news:yBEkd.11$kY3.0@newsfe3-gui.ntli.net...
>
>
> >
> > Was TAPESIZE changed as well?
>
> No
>
> >
> > How big is the disk file backup? I don't think that 9.21 allowed for
> > disk sizes greater than 2G, but this is Windows here, so . . . . ?
> >
> 216 Gb - allowed on Windows
>
Wow -things vanish fast on here.
Thanks for the suggestions - and sorry for the late thanks - I don't have
nntp at work as our NetworkMunky is a nazi that way. At home I have ntl -
anyone in the uk may know how random their access can be occasionally. Plus
I'm a bit under the weather and have spent most non -work hours in bed.
(thanks Obnox - but unfortunately no - my Fiancee's out of the continent) .
Anyway in case someone's come up with an idea - the facts were

Tapesize is about 4 Tb (to discfile) - I know that sounds odd to

<flame invitation>
the sandal wearing, ginger bearded furry toothed geek *ix bods. But I can
do that - so I do
</flame invitation>

and like I said - the backup was from a restore from live which had then
been significantly reduced in size ~ (just in case this jogs memory)
Tried a second restore from the same backup file - same problem

Blocksize was 1024 - archecker read it fine with that size set, obviously,
failed on anything else (well -when I say anything - on anything else I've
ever used) as one would expect.
I stopped the erstore with a <Ctrl> C - came up in a mess and fell over (as
we all probably expected).
Would have liked to have had a play with it - but had to bin it - put on a
backup from 2 days later - came back on fine.
Sorry I couldn't give more - because like anyone else here I would have
liked to have got to the bottom of it in case it ever happens in DR. Don't
know what's happened with Informix's stunning support - but the respose that
I got back from our intermediate support (i.e. please call IBM now, if I'm
calling I have a genuine problem) from IBM was such a joke they passed it
on without comment. So bad, I didn't even have to explain to my manager why
it was rubbish. OK - test environment - but ..... I was quite shocked.
Whatever happened to Domineque?

Andrew do you have a case number?


"Andrew" <andrew.gothard1@<nospam ta}ntlworld.com> wrote in message news:<K4bmd.214$C%4.30@newsfe6-win.ntli.net>...
> "Andrew" <andrew.gothard1@<nospam ta}ntlworld.com> wrote in message
> news:yBEkd.11$kY3.0@newsfe3-gui.ntli.net...
> >
> >
> > >
> > > Was TAPESIZE changed as well?
> >
> > No
> >
> > >
> > > How big is the disk file backup? I don't think that 9.21 allowed for
> > > disk sizes greater than 2G, but this is Windows here, so . . . . ?
> > >
> > 216 Gb - allowed on Windows
> >
> Wow -things vanish fast on here.
> Thanks for the suggestions - and sorry for the late thanks - I don't have
> nntp at work as our NetworkMunky is a nazi that way. At home I have ntl -
> anyone in the uk may know how random their access can be occasionally. Plus
> I'm a bit under the weather and have spent most non -work hours in bed.
> (thanks Obnox - but unfortunately no - my Fiancee's out of the continent) .
> Anyway in case someone's come up with an idea - the facts were
>
> Tapesize is about 4 Tb (to discfile) - I know that sounds odd to
>
> <flame invitation>
> the sandal wearing, ginger bearded furry toothed geek *ix bods. But I can
> do that - so I do
> </flame invitation>
>
> and like I said - the backup was from a restore from live which had then
> been significantly reduced in size ~ (just in case this jogs memory)
> Tried a second restore from the same backup file - same problem
>
> Blocksize was 1024 - archecker read it fine with that size set, obviously,
> failed on anything else (well -when I say anything - on anything else I've
> ever used) as one would expect.
> I stopped the erstore with a <Ctrl> C - came up in a mess and fell over (as
> we all probably expected).
> Would have liked to have had a play with it - but had to bin it - put on a
> backup from 2 days later - came back on fine.
> Sorry I couldn't give more - because like anyone else here I would have
> liked to have got to the bottom of it in case it ever happens in DR. Don't
> know what's happened with Informix's stunning support - but the respose that
> I got back from our intermediate support (i.e. please call IBM now, if I'm
> calling I have a genuine problem) from IBM was such a joke they passed it
> on without comment. So bad, I didn't even have to explain to my manager why
> it was rubbish. OK - test environment - but ..... I was quite shocked.
> Whatever happened to Domineque?

Andrew,

Did your support provider think that IBM's response was as poor as you
have suggested?

I guess you have told them its unacceptable, have they complained to
the appropriate support manager about the repsonse they have received
from IBM?




"Andrew" <andrew.gothard1@<nospam ta}ntlworld.com> wrote in message news:<K4bmd.214$C%4.30@newsfe6-win.ntli.net>...
> "Andrew" <andrew.gothard1@<nospam ta}ntlworld.com> wrote in message
> news:yBEkd.11$kY3.0@newsfe3-gui.ntli.net...
> >
> >
> > >
> > > Was TAPESIZE changed as well?
> >
> > No
> >
> > >
> > > How big is the disk file backup? I don't think that 9.21 allowed for
> > > disk sizes greater than 2G, but this is Windows here, so . . . . ?
> > >
> > 216 Gb - allowed on Windows
> >
> Wow -things vanish fast on here.
> Thanks for the suggestions - and sorry for the late thanks - I don't have
> nntp at work as our NetworkMunky is a nazi that way. At home I have ntl -
> anyone in the uk may know how random their access can be occasionally. Plus
> I'm a bit under the weather and have spent most non -work hours in bed.
> (thanks Obnox - but unfortunately no - my Fiancee's out of the continent) .
> Anyway in case someone's come up with an idea - the facts were
>
> Tapesize is about 4 Tb (to discfile) - I know that sounds odd to
>
> <flame invitation>
> the sandal wearing, ginger bearded furry toothed geek *ix bods. But I can
> do that - so I do
> </flame invitation>
>
> and like I said - the backup was from a restore from live which had then
> been significantly reduced in size ~ (just in case this jogs memory)
> Tried a second restore from the same backup file - same problem
>
> Blocksize was 1024 - archecker read it fine with that size set, obviously,
> failed on anything else (well -when I say anything - on anything else I've
> ever used) as one would expect.
> I stopped the erstore with a <Ctrl> C - came up in a mess and fell over (as
> we all probably expected).
> Would have liked to have had a play with it - but had to bin it - put on a
> backup from 2 days later - came back on fine.
> Sorry I couldn't give more - because like anyone else here I would have
> liked to have got to the bottom of it in case it ever happens in DR. Don't
> know what's happened with Informix's stunning support - but the respose that
> I got back from our intermediate support (i.e. please call IBM now, if I'm
> calling I have a genuine problem) from IBM was such a joke they passed it
> on without comment. So bad, I didn't even have to explain to my manager why
> it was rubbish. OK - test environment - but ..... I was quite shocked.
> Whatever happened to Domineque?

"scottishpoet" <dryburghj@yahoo.com> wrote in message
news:81714288.0411160302.52984814@posting.google.c om...
> Andrew,
>
> Did your support provider think that IBM's response was as poor as you
> have suggested?
>
> I guess you have told them its unacceptable, have they complained to
> the appropriate support manager about the repsonse they have received
> from IBM?
>
>
>
Don't know mate. I'm from NE England, so you probably have an idea about
how low my nonsense threshold is. Given that this was support for a test
system and the retrieval of the problem was a largely academic exercise as I
had another backup to restore from, which worked fine - I though I had
better calm down for a couple of days before responding.
I have a case number from our intermidiate support. Unfortunately we no
longer have direct access to Informix/IBM. They will no doubt have a case
number, and I'm now closer to the mood of offering constructive suggestions,
rather than the kind of bollocking that would make Alex Ferguson hide under
his desk.
I also have to make sure that they've passed on the appropriate information
in the first place, given that the reply I got was based on *ix file size
limits - in spite of the fact I repeat the OS (Windows 2000 Advanced Server)
sp and Informix version on every contact, although no-one appeared to have
read onconfig, mind.

"Andrew" <andrew.gothard1@<nospam ta}ntlworld.com> wrote in message
news:f5vmd.174$zg.130@newsfe4-gui.ntli.net...
>
> "scottishpoet" <dryburghj@yahoo.com> wrote in message
> news:81714288.0411160302.52984814@posting.google.c om...
>> Andrew,
>>
>> Did your support provider think that IBM's response was as poor as you
>> have suggested?
>>
>> I guess you have told them its unacceptable, have they complained to
>> the appropriate support manager about the repsonse they have received
>> from IBM?

Perhaps his application vendor is not too proficient at Informix itself and
either can't judge the quality of the advice it is receiving from IBM, or is
in fact receiving good advice from IBM but is unable to disseminate it
accurately?

Don't forget also this *very* important factor: if he's in the UK the
application vendor will not be talking to IBM directly (they aren't allowed
to), but will themselves be logging the call through its distributor,
probably OpenPSL, who in turn logs it with IBM and acts as a go-between for
the duration of the call.

So: end user->application vendor->distributor->IBM->distributor->application
vendor->end user. That's quite some chain. Chinese Whispers anyone ... ?