root@cancol.oz (Super User) (04/11/90)
Help please!
I have 2 dcpi servers (DN3500,4500 at sr10.2) each carrying 8 PCs
running DPCI dpcinet version 5.0. Everything works fine initially,
until dterm starts freezing up at login time, ie. at the point
before the banner message appears and the login: prompt but after
the inverted text status lines appear. NOTE: Dterm logins are
possible for a while initially, but almost every day I'm having to
kill and restart the PCI server processes and reboot all the PCs.
The only clues I have are that I can almost guarrantee causing the
problem by causing vtserver to startup (by executing "vt100" on the
console, or "more" etc). The problem also seems to coincide with
the appearance of dpci processes attaching themselves to ttyp*
devices as shown in the "ps" output below. Incidently, killing the
server processes kills all dpci_server processes except those attached
to ttyp* devices.
Also, the problem appears to arrise on both servers at about the same
time (I should add that all the PCs are in one laboratory and that
one of the servers is an SPM [no console]).
What else can I add: the PCs and 2 servers are on their own ethernet
subnetwork with the console server [DN4500] as router.
"ps" output from the DN4500.
PID TTY STAT TIME COMMAND
1 ? S < 0:37 /etc/init
2 ? R 8022:21 null
3 ? S 5:44 purifier
4 ? S 0:07 purifier
5 ? S 6:19 unwired_dxm
6 ? S 0:00 pinger
7 ? S 0:00 netreceive
8 ? S 0:40 netpaging
9 ? S 2:05 wired_dxm
10 ? S 4:46 netrequest
13 ? S 0:02 netreceive
14 ? S 0:58 router
24 ? S 0:40 netpaging
25 ? S 0:40 netpaging
26 ? S 4:46 netrequest
27 ? S 4:46 netrequest
784 ? S 8:19 /etc/tcpd
790 ? S 1:22 /etc/routed -f
793 ? S 0:00 /etc/inetd
796 ? S 0:19 /etc/timed -M
799 ? S 0:00 /etc/ncs/llbd
805 ? S 0:06 /usr/lib/sendmail -bd -q1h
811 ? S 0:07 /sys/spm/spm
815 ? S 0:01 /sys/mbx/mbx_helper
817 ? S 3:12 /sys/hardcopy/prmgr -c /sys/hardcopy/prmgr_config
820 ? S 4:19 /sys/hardcopy/prsvr /sys/hardcopy/prsvr_config
822 ? S 3:45 /sys/hardcopy/pre10q -site //delta -n pre10q
824 ? S 0:00 /etc/writed
830 ? S < 0:27 /etc/Xapollo -K /usr/X11/lib/keyboard/keyboard.config
-D1 s+r-
833 ? S < 5:29 dm
1200 ? S 0:00 dpci -netbios dpcinet -pingtime 5
1202 ? S 5:43 /sys/dpci/daemons/dpcinetd -pingtime 5
1211 ? S 0:31 dpci_server -netbios dpcinet -pingtime 5 -signal
1227 ? S 0:09 dpci_server -netbios dpcinet -pingtime 5 -signal
1229 ? S 0:09 dpci_server -netbios dpcinet -pingtime 5 -signal
1230 ? S 0:09 dpci_server -netbios dpcinet -pingtime 5 -signal
1277 ? S 0:17 dpci_server -netbios dpcinet -pingtime 5 -signal
1279 ? S 0:18 dpci_server -netbios dpcinet -pingtime 5 -signal
1281 ? S 0:15 dpci_server -netbios dpcinet -pingtime 5 -signal
1282 ? S 0:11 dpci_server -netbios dpcinet -pingtime 5 -signal
1298 ? S 0:00 dpci_server -netbios dpcinet -pingtime 5 -signal
1330 ? R 0:00 rshd
1331 ? R 0:00 csh -c ps agx
1332 ? R 0:00 ps agx
1280 ttyp0 S 0:10 dpci_server -netbios dpcinet -pingtime 5 -signal
1283 ttyp0 S 0:00 /sys/dpci/dterm_domain /sys/dm/login_sh 1280 151f8
N dpcinet 0
We invariably start dterm with the command line
'dterm -program "/sys/dm/login_sh" -login'
This may be asking for trouble but: the PCs are often rebooted without
closing down the PCI connections. This is unavoidable in our student
environment.
Thanks for your time,
+----------------------+---+
| Ross Johnson |:-(| ACSnet: ross@cancol.oz.au
| Info. Sciences & Eng.|___| ARPA: ross%cancol.oz.au@uunet.uu.net
| University of Canberra | UUCP: {uunet,ukc}!munnari!cancol.oz.au!ross
| PO Box 1 | CSNET: ross%cancol.oz@australia
| Belconnen ACT 2616 | JANET: ross%au.oz.cancol@EAN-RELAY
| AUSTRALIA | BITNET: ross%cancol.oz.au@relay.cs.net
+--------------------------+ross@cancol.oz (Ross Johnson) (04/12/90)
In article <344@cancol.oz>, root@cancol.oz (That's me) writes: > Help please! > > I have 2 dcpi servers (DN3500,4500 at sr10.2) each carrying 8 PCs > running DPCI dpcinet version 5.0. Everything works fine initially, > until dterm starts freezing up at login time, ie. at the point > before the banner message appears and the login: prompt but after > the inverted text status lines appear. NOTE: Dterm logins are > possible for a while initially, but almost every day I'm having to > kill and restart the PCI server processes and reboot all the PCs. > Another clue: I find the following 3 entries in the /usr/adm/dpci_error_log file (there are others which are identical). Can anyone interpret the the last entry, what does it mean "creating a temporary pty"? ******************************************************************************** Entry created on 11-Apr-1990 09:40:51 Node id: 1CF47 Process name: dpcinet_server_15201 DOMAIN(R)/PCI(TM) DTERM 5.00 October 31, 1989 3:07:58 pm ******************************************************************************** Entry created on 11-Apr-1990 09:40:51 Node id: 1CF47 Process name: dpcinet_server_15201 SERVER/DTERM: PC DTERM revision 5.00. ******************************************************************************** Entry created on 11-Apr-1990 09:40:52 Node id: 1CF47 Process name: dpcinet_server_15201 SERVER/DTERM: Couldn't open a pty device, creating a temporary pty - name not found (stream manager/IOS) ******************************************************************************** Thanks again, Perplexed