List of programs to load -- SHM.TO.LOAD -- does not exist.

RelianceDss · Post by **RelianceDss** » Thu Nov 01, 2007 12:58 am

Hi friends,

We are getting following error while bringing up the datastage.
List of programs to load -- SHM.TO.LOAD -- does not exist.

When we try to connect using client it gives following error:
Failed to connect to host: (Cannot find UVRPC in host services file (81014))

Ur suggestions will be highly appreciated.
Let me know if any more input is required from our side.

Thnk and rgrds
deepak

ray.wurlod · Post by **ray.wurlod** » Thu Nov 01, 2007 1:03 am

Someone at your site has been playing around with "UniVerse" things and without knowledge of what they're doing.

The SHM.TO.LOAD file exists if an Administrator has identified BASIC routines that are to be pre-loaded into shared memory so that the one copy of object code can be used by all process, rather than each loading their own copy into shared memory. Clearly someone has only partially completed setup of this feature.

The message about uvrpc is because someone is trying to connect using a UniVerse client (perhaps one written in DataStage version 5.2 or earlier, which is possible at Reliance). You will not get this message from DataStage clients in version 7x, because they use dsrpc rather than uvrpc.

Stop trying to do UniVerse things is the solution. They are no longer (since version 6.0) guaranteed to work.

If you genuinely run UniVerse database (which, from my memory, Reliance does not), then you do need an entry for the UniVerse RPC process in /etc/hosts, which shows it listening on TCP port number 31438.

RelianceDss · Post by **RelianceDss** » Thu Nov 01, 2007 1:25 am

Hi Ray,

Thnx for ur prompt reply.

We are using DS7.5.2 here. As for routines, we have some basic routines used in jobs, but dont now if they get loaded in shared memory or not. We will try to investigate the universe thing for sure but do we have any solution for making this work. We are not able to connect to the server right now.
We have a separate quality environment also which is running fine. It is almost the copy of corrupt environment. Can that be of any help, in case?

thnx and rgrds

ray.wurlod · Post by **ray.wurlod** » Thu Nov 01, 2007 1:43 am

You may be using 7.5.2 now, but your site has had DataStage for many years, and there may be some "leftovers" from previous version practice.

Inspect the SHM.TO.LOAD file on the system that is working. You will see that each routine is represented by its account (project), directory file name and program file name.

Also on the working system, execute the command analyze.shm -b to determine whether BASIC shared memory is actually being used.

RelianceDss · Post by **RelianceDss** » Thu Nov 01, 2007 4:17 am

ray.wurlod wrote:You may be using 7.5.2 now, but your site has had DataStage for many years, and there may be some "leftovers" from previous version practice.

Inspect the SHM.TO.LOAD file on the system that is working. You will see that each routine is represented by its account (project), directory file name and program file name.

Also on the working system, execute the command analyze.shm -b to determine whether BASIC shared memory is actually being used.

This is a new box and we started with 7.5.2 afresh. There was no previous installation on this box. analyze.shm -b didnt give any output.
We are able to bring up/down the engine with admin user but the client can't connect. It gives "(Cannot find UVRPC in host services file (81014))" error.
When we bring up the engine from root it works absolutely fine.
Plz suggest how can we make it work from admin user. It was working ok till recently.
thnx and rgrds

ArndW · Post by **ArndW** » Thu Nov 01, 2007 6:06 pm

can the dsadm user read the /etc/services file? If yes, is there an entry for dsrpc in it (should be at the end)

RelianceDss · Post by **RelianceDss** » Fri Nov 02, 2007 1:32 am

ArndW wrote:can the dsadm user read the /etc/services file? If yes, is there an entry for dsrpc in it (should be at the end)

Yes and yes.

Also when we bring up the engine with admin user it asks for passwd which it never used to before. With root the engine comes up fine.

Since we have spent quite a lot of time on this and also many posts in this forum has suggested reinstall, there are a few questions if we go for reinstall:
1. Should we directly reinstall on the current setup or shall we first formatt everything and then install afresh? Best practice?
2. Should we use admin user or root to install DS? Best practice?
3. What else should we planout if we go for a reinstall?

Thns and Rgrds
deepak

ray.wurlod · Post by **ray.wurlod** » Fri Nov 02, 2007 2:36 am

Look at the /etc/services file for the system that is working. Is there an entry there for uvrpc or uvrpcd? If so, mimic it in the /etc/services file of the system where it is not working. Otherwise duplicate the dsrpc (or dsrpcd) entry in the /etc/services file in the new system, substituting "uv" for "ds" in the name of the service.

DataStage clients post 6.0 simply should not be looking for uvrpc - they should be looking for dsrpc. Report this to your support provider.

RelianceDss · Post by **RelianceDss** » Fri Nov 02, 2007 2:53 am

/etc/services file has dsrpc entry in both the systems and they are exactly the same. At times we also get incorrect/username and password error when logging on with the client.

thnx & rgrds

RelianceDss · Post by **RelianceDss** » Mon Nov 19, 2007 6:45 am

We went for a update. It solved the issue.