Hi,

after checking the feasibilities I decided to jump into real Proview programming and I now encounter configuration problem which is probably beginner one but that I don't understand.

I started with a fresh new install on ubuntu 10.10 with proview (.deb) 4.8.0 (installation problem is now only memories.... good job guys !!) and I followed getting starting guide.

1) Everything worked fine until graphical editor design. First I need to configure my graph as FastAvail1 because FastAvail0 doesn't appear on the banner of the operator station. Is that normal ?

2) After that I tried to modify my plc program to discover other modules and I was unable to add modules (I tried toggle editor mode on and off without success) and I don't understand why. I then decided to stop proview and to restart it. and I got the following error in the main console:

/usr/local/pwrp/gettingstarted/src/db/directory.wb_load:2:16: unexpected token: 01
E Wbl parse error: /usr/local/pwrp/gettingstarted/src/db/directory.wb_load line: 2
F Wbl terminated
%LDH-E-NOSUHVOL, no such volume exists
Unable to open journal file $pwrp_tmp/nameless_1557.gjl
**Warning, systemname not found


I looked in the directory.wb_load file and the first lines are the following:
Volume directory $DirectoryVolume 254.254.254.253
Body SyBody 01-JANv.-1970 01:00:00:00
EndBody
.....

Note: if 2:16 is the line/column reference, it indicates the \"01\" of JANv 1970 date.

I wonder why the date is stucked at the pc default one ? (is that normal as I didn't looked inside this file before).



3) Another more generally question: is there a document explaining the philosophy of the Volume / node / process station decomposition. I must confess that I don't understand well enough to be able to move from the getting started guide conf to my own needs (5 process station, 3 global monitoring stations + 1 storage). Should I create one node per physical machine with each of them having their volume ?).


Thanks in advance for answering all these questions.

Snarf77

Hi Snarf,

1) No this is not normal, but I'm not able to reproduce this behavior.

2) This is the problem with the French locale. The date should be written 1-JAN-1970. We obviously haven't fixed this problem yet, so I'm afraid you will have to change to English.

3) A node represents a physical machine, and for each node you need a RootVolume, so you'll need 8 NodeConfig and 8 RootVolumeConfig objects configured in the directory volume. For the storage station you need a SevNodeConfig, and also a RootVolume, but this should not be configured with a RootVolumeConfig (it doesn't contain any objects, it just works as a network address for the storage station).

It is also possible divide the nodes into several projects. Some developers prefer to have one project for each process station and a common project for operator stations.

/Claes

Hi Claes,

A big thank for your very useful answer (as usual).

I'll change to english to avoid this bug. Indeed I'm not in a hurry for this bug to be fixed but I you could just inform me when it will be corrected, I would be grateful (mainly for my colleagues with (more) limited knowledge in english)

When I have changed to english I will check if I'm still limited with fastavail0. In this case I will try to provide more information for your team to have a look.

Concerning the decomposition of the volumes, I just have an additionnal question: if all the process station are identical (same software and I/O to control similar drives) is it more interested to have one common project for all process station or keep them separated . What are the advantage / drawback of one solution against the other ?

Sorry for the inconvenience and thank again for your time

Snarf

Hi Snarf,

If you have 24/7 production it is usually easier to handle upgrades of separate projects than of one large project. If you have common resources like classvolumes, process graphs etc, one common project is a better choice, so I think your on the right track with a single project.

/Claes