[HECnet] Multinet 5.5 issues on VMS 7.3 (VAX) under SIMH
Keith Halewood
Keith.Halewood at pitbulluk.org
Fri Jun 28 15:40:21 PDT 2019
Hi,
Just FYI more than anything.
I think, for my purposes, that NTP is a bit of an overkill. Under SIMH, I've noticed that the TODR works well enough for the range of simulations I use, except the newer vax8200 simulator.
I was considering just using MULTINET RDATE and leaving my hosts in the UTC timezone. However, what I've done is implemented a version of 'date' (port 37) on the host linux system that returns a 32 bit number of seconds since 1st Jan 1900, offset by the local time. For me, that's GMT <-> BST. It's good enough for me.
Multinet's network interface is a bit of a mish-mash for my mind. I have a recollection of using CMUTEK TCP/IP under VAX/VMS several 'centuries' ago and that was a pure $QIO(W)-based interface, wonderfully straightforward. Multinet offers a $QIO interface... and then ruins it by insisting on socket-like data structures for establishing connections - yuck.
It seems I'm getting really intolerant in my old age.
Keith
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On Behalf Of Keith Halewood
Sent: 23 June 2019 18:48
To: hecnet at Update.UU.SE
Subject: RE: [HECnet] Multinet 5.5 issues on VMS 7.3 (VAX) under SIMH
Hi Supratim,
No crash! So I imagine I just have to use 'pool' correctly to the Europe ntp pool probably. Thanks for your help. I was beginning to doubt SIMH, especially with my problems with unzip'ing from various DECUS sources. I even removed some... erm... termination dates... just in case a VMS restart resulted in an inconvenient clock time.
Keith
From: owner-hecnet at Update.UU.SE<mailto:owner-hecnet at Update.UU.SE> [mailto:owner-hecnet at Update.UU.SE] On Behalf Of Supratim Sanyal
Sent: 23 June 2019 17:26
To: hecnet at Update.UU.SE<mailto:hecnet at Update.UU.SE>
Subject: Re: [HECnet] Multinet 5.5 issues on VMS 7.3 (VAX) under SIMH
Keith,
I pasted your NTP.CONF file in and got the exact same crash as you. Then I tried the generic NIST servers (time-a-g.nist.gov,time-b-g.nist.gov,time-c-g.nist.gov,time-d-g.nist.gov) and also got the same crash. BUT the following works - can you try it?
If it works, my guess is Multinet's NTPD has a problem with servers that forward to other servers but are not pool servers. Weird, I didn't expect this.
driftfile MULTINET:NTP.DRIFT
statistics loopstats peerstats clockstats
filegen loopstats file loopstats type day enable
filegen peerstats file peerstats type day enable
filegen clockstats file clockstats type day enable
server tick.usno.navy.mil iburst prefer
server tock.usno.navy.mil iburst prefer
restrict -4 default kod notrap nomodify nopeer noquery limited
restrict -6 default kod notrap nomodify nopeer noquery limited
restrict 127.0.0.1
restrict ::1
restrict source notrap nomodify noquery
On 06/23/2019 10:29 AM, Keith Halewood wrote:
Hi Supratim,
I just added server entries to google so my ntp.conf is initially:
driftfile MULTINET:NTP.DRIFT
statistics loopstats peerstats clockstats
filegen loopstats file loopstats type day enable
filegen peerstats file peerstats type day enable
filegen clockstats file clockstats type day enable
server time1.google.com burst iburst prefer
server time2.google.com burst iburst prefer
server time3.google.com burst iburst prefer
server time4.google.com burst iburst prefer
restrict -4 default kod notrap nomodify nopeer noquery limited
restrict -6 default kod notrap nomodify nopeer noquery limited
restrict 127.0.0.1
restrict ::1
restrict source notrap nomodify noquery
When I start NTP, I eventually see this in the logfile:
$ Set NoOn
$ VERIFY = F$VERIFY(F$TRNLNM("SYLOGIN_VERIFY"))
23 Jun 15:24:02 [multinet[537]: ntpdc 4.2.8p4 at 1.3265-o<mailto:4.2.8p4 at 1.3265-o> Mon Nov 2 19:53:28 UTC 2
015 (1): Starting
23 Jun 15:24:02 [multinet[537]: Command line: ix$dua0:[multinet.ix.syscommon.][m
ultinet]ntpd.exe;2
23 Jun 15:24:08 [multinet[537]: proto: precision = 10000.000 usec (-7)
23 Jun 15:24:08 [multinet[537]: proto: fuzz beneath 140.845 usec
23 Jun 15:24:08 [multinet[537]: Listen and drop on 0 v6wildcard [::]:123
23 Jun 15:24:08 [multinet[537]: Listen and drop on 1 v4wildcard 0.0.0.0:123
23 Jun 15:24:08 [multinet[537]: Listen normally on 2 se0 192.168.2.36:123
23 Jun 15:24:08 [multinet[537]: Listen normally on 3 lo0 127.0.0.1:123
The command 'multinet ntpq -p' gives me:
$ multinet ntpq -p
remote refid st t when poll reach delay offset jitter
==============================================================================
DISK$MULTINETBLD:[MULTINET-SRC.MULTINET.NTP4.LIBNTP]DECODENETNUM.C;2:78: INSIST(
ai->ai_addrlen <= sizeof(*netnum)) failed.
%SYSTEM-F-OPCCUS, opcode reserved to customer fault at PC=000856D9, PSL=03C00000
Improperly handled condition, image exit forced.
....and a signal, stack and register dump.
So... I'm a bit stuck still. Any pointers (preferably not null ones :) ) gratefully received.
Keith
From: owner-hecnet at Update.UU.SE<mailto:owner-hecnet at Update.UU.SE> [mailto:owner-hecnet at Update.UU.SE] On Behalf Of Supratim Sanyal
Sent: 23 June 2019 12:56
To: hecnet at Update.UU.SE<mailto:hecnet at Update.UU.SE>
Subject: Re: [HECnet] Multinet 5.5 issues on VMS 7.3 (VAX) under SIMH
Hi Keith,
On 06/23/2019 05:46 AM, Keith Halewood wrote:
Hi,
Is anybody running the NPTD functionality in Multinet 5.5 on VAX VMS 7.3 (under SIMH)? If so, how did you get it to work. I switched it on with a simple NTP.CONF and tried to query it. The logfile doesn't change much and I/O counts go up slowly. My attempts to query it usually result in a traceback. I'm a bit concerned by the lack of an NTPDATE command mentioned in the documentation, despite having done a full install. The timezone facility is a bit primitive too.
Works okay on multiple SIMH OpenVMS/VAX 7.3 instances. Here is what I get:
$ multinet ntpq -p
remote refid st t when poll reach delay offset jitter
==============================================================================
*104.167.106.93 128.233.150.93 2 u 11 64 1 49.723 267.931 7.813
pool.ntp.org .POOL. 16 p - 64 0 0.000 0.000 7.813
+sanyalnet-cloud 45.73.0.50 2 u 15 64 5 39.625 18.945 12.511
+sanyalnet-cloud 198.166.1.59 2 u 83 64 6 39.535 60.975 51.664
+sanyalnet-cloud 142.3.100.2 2 u 141 64 4 43.671 40.799 23.688
+vps5.ctyme.com 216.218.254.202 2 u 10 64 7 119.906 -27.318 46.828
+tick.eoni.com 216.228.192.53 2 u 7 64 7 79.618 34.703 28.908
+horp-bsd01.horp 146.186.222.14 2 u 8 64 7 29.949 23.332 36.004
+meanwhile.clfs. 200.98.196.212 2 u 12 64 7 20.000 -3.831 43.376
And here is my MULTINET:NTP.CONF; 10.42.0.0/24, 10.100.0.0/24 and 10.200.0.0/24 entries are my local subnets with unrestricted access. Also I run my own public NTP servers, but any good public servers should work. You can just have the pool entry if you want.
$ type multinet:ntp.conf
driftfile MULTINET:NTP.DRIFT
statistics loopstats peerstats clockstats
filegen loopstats file loopstats type day enable
filegen peerstats file peerstats type day enable
filegen clockstats file clockstats type day enable
server sanyalnet-cloud-vps.freeddns.org burst iburst prefer
server sanyalnet-cloud-vps2.freeddns.org burst iburst prefer
server sanyalnet-cloud-vps3.freeddns.org burst iburst prefer
server sanyalnet-cloud-vps4.duckdns.org burst iburst prefer
pool pool.ntp.org iburst
restrict -4 default kod notrap nomodify nopeer noquery limited
restrict -6 default kod notrap nomodify nopeer noquery limited
restrict 127.0.0.1
restrict ::1
restrict source notrap nomodify noquery
restrict 10.42.2.0 mask 255.255.255.0 notrust
restrict 10.100.0.0 mask 255.255.255.0 notrust
restrict 10.200.0.0 mask 255.255.255.0 notrust
I'm thinking of giving up with it. I have local TCP services on the machines hosting SIMH instances. One of them returns a local time string in VMS format, so I may make use of that in an RDATE-like fashion.
Keith
Hope this is of some help.
Best,
Supratim
--
Sent via Thunderbird/MX-Linux on an overheated Compaq Presario CQ61
--
Sent via Thunderbird/MX-Linux on an overheated Compaq Presario CQ61
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sonic.net/pipermail/hecnet-list/attachments/20190628/4e575b15/attachment-0001.html>
More information about the Hecnet-list
mailing list