I support an older application written in Clarion 4 based on Pervasive
8.5 btrieve files. One of my customers is using a Citrix server farm
connecting to a SAN through a Windows 2000 Advanced server acting as
the "NAS". The customer is using the Workgroup Engine and it is
installed on each of their 74 Citrix servers. They are not setting a
permanent gateway for redundancy/load balancing reasons. They are
complaining about a couple of issues. The main issue is that they say
that periodically they will "lose" hundreds of records and then
mysteriously after a period of time, they will return. I haven't
actually been able to witness the loss of records, but I just wanted to
see if anyone else has heard of any similar issue and, if so, what they
did to resolve it.

Second issue is that due to the fact that they do not set a permanent
gateway and they are using the Citrix servers in a load balancing
configuration, their users could be connecting to the data files
through several different Citrix servers even though only one is
actually acting as the gateway. From time to time, they will have
trouble acessing the data files from the Clarion application, and we
have found that if we use the Pervasive Monitor utility we see several
connections to the data files from each of these Citrix servers even
though they have temporarily disabled the users from connecting to the
Citrix servers. My question is what would cause the Citrix servers to
maintain their connections to the gateway workgroup engine and is there
anything that can be configured to alleviate this problem. Currently,
the only way to "fix" it is to clear all users from within the Monitor
utility.

Upgrading to the Enterprise Engine is not an option at this time due to
licensing costs.

Thanks in advance.

The WGE is working with 74 different Citrix boxes??????? I'm not even
sure which license model this would use, although I could suspect that
it could work if very few users ever share access to the same database
files.

Normally, the WGE handles a MAXIMUM of 5 concurrent users in a
workgroup. I can easily see a WGE failing temporarily and having
another take over for it, causing records to disappear and reappear
from time to time. This is definitely beyond what this engine was
designed for, and this configuration is NOT recommended (and probably
not supported by Pervasive, either).

You could try patching to PSQLV8.6 to see if that helps...

The only real solution in this case is to switch to the Client/Server
Engine, which would require a single dedicated server with appropriate
licensing for the number of users attaching to that server. This would
definitely free up a lot of the problems, and it should stabilize the
entire configuration.
Goldstar Software Inc.
Building on Btrieve(R) for the Future(SM)
Bill Bach
BillBach (AT) goldstarsoftware (DOT) com
http://www.goldstarsoftware.com
*** Chicago: Pervasive.SQL Service & Support - July 19-22 ***
*** Chicago: Pervasive DataExchange Class - July 27-29 ***

Dinger wrote:

Well, it's a long story as to WHY and HOW we are running such a large #
of workgroup engines in this environment. Suffice it to say that our
CIO had a business relationship with someone at Pervasive and was able
to negotiate a very inexpensive deal with them enabling our company to
buy licenses for all of our current customers users rather than pass
the cost onto them. Data corruption with our previous database was
reaching critical levels and this prompted the switch to a more stable
database engine. I hate that we were forced to go this route but I
still have to support the "solution".

Anyway, I have been informed that switching this customer to the
Enterprise engine is not an option at this time...but for arguments
sake...how many concurrent connections could the Enterprise Engine
handle? This particular customer has approximately 60 unique databases
stored on a SAN which are accessed through the 74 Citrix servers via a
Windows 2000 Advance Server acting as a "NAS". They have a gigabit
connection between the Citrix servers and the NAS and a fiber channel
connection from the NAS to the SAN. Each database could possibly
experience on average about 40-50 concurrent connections with some
exceeding 100+ users at times of peak usage. So if I do the math that
is approximately 2700+ concurrent connections going through one
Enterprise Engine to 60 different databases. Again this # could be much
lower most of the time and it also could be much higer at peak times.
Would one engine be able to handle that many connections or would
several engines on several NAS servers be required?

Technically, the Unlimited Engine should support gobs & gobs of
connections. In practice, my largest customers top out at around 1200
users on a single server.

It is all, of course, relative -- what performance do you want? How
much hardware do you have? If you need more performance than what you
can squeak out of existing hardware, then adding another server would
be advantageous. My splitting the data onto multiple servers, you have
almost infinite scalability within the performance guidxelines that you
set.
Goldstar Software Inc.
Building on Btrieve(R) for the Future(SM)
Bill Bach
BillBach (AT) goldstarsoftware (DOT) com
http://www.goldstarsoftware.com
*** Chicago: Pervasive.SQL Service & Support - July 19-22 ***
*** Chicago: Pervasive DataExchange Class - July 27-29 ***

Dinger wrote: