Icon View Thread

The following is the text of the current message along with any replies.
Messages 1 to 10 of 12 total
Thread C/S and index errors
Thu, Nov 29 2007 12:55 AMPermanent Link

"Al Vas"
Hi,
I just want to revisit this.  We have a number of clients who use one of our
products.  All of them run in client/server mode and it is quite a high
volume function in terms of records created.  More than I like, we have
index mismatches which require a rebuild of the table.  It is usually, but
not always, the table that is accessed the most frequently by many users.

Is it probable that our own coding is causing this in some way?  As far as I
understand in client/server there is very little, other than a server
shutdown, or abnormal database engine shutdown that would cause an index
error.  If it is potentially coding, what should we be looking for?

Thanks

Alex

Thu, Nov 29 2007 3:18 AMPermanent Link

Roy Lambert

NLH Associates

Team Elevate Team Elevate

Al


What version and what's the structure of the table? Are there memo fields being updated?

Roy Lambert
Thu, Nov 29 2007 4:14 PMPermanent Link

Tim Young [Elevate Software]

Elevate Software, Inc.

Avatar

Email timyoung@elevatesoft.com

Alex,

<< I just want to revisit this.  We have a number of clients who use one of
our products.  All of them run in client/server mode and it is quite a high
volume function in terms of records created.  More than I like, we have
index mismatches which require a rebuild of the table.  It is usually, but
not always, the table that is accessed the most frequently by many users.

Is it probable that our own coding is causing this in some way?  As far as
I understand in client/server there is very little, other than a server
shutdown, or abnormal database engine shutdown that would cause an index
error.  If it is potentially coding, what should we be looking for? >>

Is this 3.x ?  If so, then I would suggest that you upgrade your customers
to 4.x at the very least.  3.x simply is not going to have any more fixes,
so there isn't much I can do to help you.

--
Tim Young
Elevate Software
www.elevatesoft.com

Thu, Nov 29 2007 8:01 PMPermanent Link

"Al Vas"
Hi Roy,

It is V3.30.  Many tables do include memo fields.  The highly used table has
1 memo field and 14 secondary indexes

Alex

"Roy Lambert" <roy.lambert@skynet.co.uk> wrote in message
news:48944BF9-E209-476D-8E0F-98B0A5D7522A@news.elevatesoft.com...
> Al
>
>
> What version and what's the structure of the table? Are there memo fields
> being updated?
>
> Roy Lambert
>

Thu, Nov 29 2007 8:05 PMPermanent Link

"Al Vas"
Hi Tim,

Yes unfortunately we just havent had the window of opportunity amongst alot
of development to upgrade to 4.  Based on our previous experience of moving
from 1 to 2 and 2 to 3 we know there is alot of work and testing and
updating of tools that we have.  I wasn't really asking for fixes to 3.x but
would be appreciative if there is any knowledge out there of any issues that
would cause these index problems.  If it is our code, I'm just interested on
what sort of coding would cause index issues as we don't do anything tricky
when it comes to table access.

Alex


"Tim Young [Elevate Software]" <timyoung@elevatesoft.com> wrote in message
news:A74ABC97-6971-4518-B96C-65D770B3B1DA@news.elevatesoft.com...
> Alex,
>
> << I just want to revisit this.  We have a number of clients who use one
> of our products.  All of them run in client/server mode and it is quite a
> high volume function in terms of records created.  More than I like, we
> have index mismatches which require a rebuild of the table.  It is
> usually, but not always, the table that is accessed the most frequently by
> many users.
>
> Is it probable that our own coding is causing this in some way?  As far as
> I understand in client/server there is very little, other than a server
> shutdown, or abnormal database engine shutdown that would cause an index
> error.  If it is potentially coding, what should we be looking for? >>
>
> Is this 3.x ?  If so, then I would suggest that you upgrade your customers
> to 4.x at the very least.  3.x simply is not going to have any more fixes,
> so there isn't much I can do to help you.
>
> --
> Tim Young
> Elevate Software
> www.elevatesoft.com
>

Thu, Nov 29 2007 10:31 PMPermanent Link

"Robert"

"Al Vas" <forgetit@forgetit.com> wrote in message
news:785F1627-08D0-4E59-B517-D3F6C3AA9150@news.elevatesoft.com...
> Hi Tim,
>
> Yes unfortunately we just havent had the window of opportunity amongst
> alot of development to upgrade to 4.  Based on our previous experience of
> moving from 1 to 2 and 2 to 3 we know there is alot of work and testing
> and updating of tools that we have.  I wasn't really asking for fixes to
> 3.x but would be appreciative if there is any knowledge out there of any
> issues that would cause these index problems.  If it is our code, I'm just
> interested on what sort of coding would cause index issues as we don't do
> anything tricky when it comes to table access.
>

There was a bug in 3.30 that was eating my lunch, corrupting indexes on
tables that had a lot of multi-user activity, finally decided to upgrade o
V4. You are right that the upgrade is not trivial, especially if you use a
lot of SQL.

If you have the V3 source, I think Tim has the fix that you can apply.

Robert


Fri, Nov 30 2007 3:16 AMPermanent Link

Roy Lambert

NLH Associates

Team Elevate Team Elevate

Al


In my case it was a combination of threads (much the same as multiple users) and full text indexing that I think caused the problem. It was eventually sorted when I was able to produce a test case. I can't remember the version though.

Roy Lambert
Fri, Nov 30 2007 2:57 PMPermanent Link

Tim Young [Elevate Software]

Elevate Software, Inc.

Avatar

Email timyoung@elevatesoft.com

Alex,

<< Yes unfortunately we just havent had the window of opportunity amongst
alot of development to upgrade to 4.  Based on our previous experience of
moving from 1 to 2 and 2 to 3 we know there is alot of work and testing and
updating of tools that we have.  I wasn't really asking for fixes to 3.x but
would be appreciative if there is any knowledge out there of any issues that
would cause these index problems.  If it is our code, I'm just interested on
what sort of coding would cause index issues as we don't do anything tricky
when it comes to table access. >>

Yes, but you have to remember that the last time I even touched 3.x was
years ago.  I really can't even remember anything relevant to the discussion
anymore at this point.

--
Tim Young
Elevate Software
www.elevatesoft.com

Sat, Dec 1 2007 3:53 AMPermanent Link

"Al Vas"
Fair enough Tim.  I guess though, if index issues are caused by our code,
then it will be no different if we upgrade to 4, so it might be wise for us
to start investigating before we move to 4.?  Has anyone had experiences
with indexing issues caused by their code in Version 4?

Regards

Alex

"Tim Young [Elevate Software]" <timyoung@elevatesoft.com> wrote in message
news:BD05B7F8-3B15-420F-9C35-F868044A5AE3@news.elevatesoft.com...
> Alex,
>
> << Yes unfortunately we just havent had the window of opportunity amongst
> alot of development to upgrade to 4.  Based on our previous experience of
> moving from 1 to 2 and 2 to 3 we know there is alot of work and testing
> and
> updating of tools that we have.  I wasn't really asking for fixes to 3.x
> but would be appreciative if there is any knowledge out there of any
> issues that would cause these index problems.  If it is our code, I'm just
> interested on
> what sort of coding would cause index issues as we don't do anything
> tricky when it comes to table access. >>
>
> Yes, but you have to remember that the last time I even touched 3.x was
> years ago.  I really can't even remember anything relevant to the
> discussion anymore at this point.
>
> --
> Tim Young
> Elevate Software
> www.elevatesoft.com
>

Sat, Dec 1 2007 12:24 PMPermanent Link

Eryk Bottomley
Al,

> Fair enough Tim.  I guess though, if index issues are caused by our code,
> then it will be no different if we upgrade to 4,

No, because the only way that your code could cause index corruption
would be if you added code to the server which caused a catastrophic
failure of the entire process (and I suspect you would know if that were
happening).

Client side code could only cause index corruption if it managed to
unearth an otherwise unknown bug in the V3.x engine and it is unlikely
that such a bug would have managed to persist through the entire product
lifetimes of V3.x AND V4.x undetected.

> to start investigating before we move to 4.?  Has anyone had experiences
> with indexing issues caused by their code in Version 4?

Only server side code causing the process to AV and die unexpectedly.

Eryk
Page 1 of 2Next Page »
Jump to Page:  1 2
Image