Dear colleagues,
I got two replies to my request. Unfotunately, they did not
solve our problem.
However, many thanks go to:
Birger A. Wathne - Birger.Wathne@vest.sdata.no
Roi Gift - gift@sqi.com
Responses follow after original posting.
--------Original Posting---------------------------------------------------
Dear colleagues,
sometimes we have problems with our network, i.e. our PCs do not have
access to the network.
The network consists of five Sparc stations running Solaris 2.3. Two of them
are servers (data and programs) for the PCs. The communication with the PCs
is established via the rpc.pcnfsd-daemon on the Sparcs and PCTCP-Software on
the PCs. One of the PCs is running PCTCP 2.3, the others (nine) are runnining
PCTCP version 2.2. The physical link is the thin wire Ethernet.
The problem is that sometimes one or two PCs cannot access to the data on the
Suns while simultaneously the other PCs can. This interuption may last several
minutes up to one hour. Both kind of PCs (PCTCP 2.2 and 2.3) are affected by
the problem.
All this sounds like a overload problem but not in all cases when the access
problem occurs all PCs are connected with the net, respectively the servers
do not run many jobs/processes.
My questions: 1) Does anybody have experiences with the same net configuration
(and the same problem) ?
2) Where is the problem located: Sun or PC ?
3) Is there any parameter which should set to a particular value ?
4) Any ideas ?
--------End of Original Posting---------------------------------------------
Possible solutions:
-------------------
The first response came from Birger who stated that the SQE test switch on
the transceiver should be turned off. This sounded very promising since I
detected that one of the switches of the PC-servers was turned off. I
contacted the SUN-hotline. They gave the same recommendation. However,
switching on the heartbeat did not abolish the problem.
The second response from Roi recommended to check whether all thin-wire
segments are properly terminated (50 ohms) and not damaged. It did not
help as well.
Once again: Many thanks to the repliers !
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:09:16 CDT