| Title: | DECnet/OSI for OpenVMS |
| Moderator: | TUXEDO::FONSECA |
| Created: | Thu Feb 21 1991 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 3990 |
| Total number of notes: | 19027 |
Config: VMS VAX 7.0
OSI 7.0
My customer has problems with dialup connections from Phase IV Clients (PC's).
When the connection drops ( PC switched off or physical link problem) he has to
wait several minutes for a MSG$PATHLOST . A SYS$DASSGN of the link before the
MSG$PATHLOST comes through results in a hang of the call.
Customer tried to get around with an asynchronous QIO-DEACCESS!SYNC followed
by SYS$DASSGN. In some cases the QIO still remains with an outstanding AST.
The hanging logical links can still be found back ( session control port and
nsp port). The links have an estimated roundtrip delay of 5000 and more.
tracing in the routers (WR250) does not reveal any trace of the link: they
clean up correctly.
Up to now only reboot helps.
Can someone assist on this?
Thx
Erwin
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 3973.1 | Sounds like DECnet-PLUS is trying harder... | TWICK::PETTENGILL | mulp | Wed May 28 1997 23:26 | 10 |
This sounds like one of the complaints with TCP/IP: "the client was powered off three days ago; why is the application server still taking up memory for that client?" Of course, the counter complaint is "why does DECnet disconnect after a few minutes? With TCP/IP I can disconnect the network for a couple of hours and then reconnect and resume right where I left off." The only thing that really seems like a bug is that the I/O isn't cleaned up faster when an abort is done. | |||||