I've now refreshed the VAX satellites system drive and installed it in the ALPHA server. The one problem I have remaining is that the VOTES the satellite is contributing to the cluster is 1. I believe for a proper satellite this should be 0.
Is this a case of updating the MODPARAMS.DAT on the satellite and autogen and reboot? Do I need to do anything with the ALPHA servers configuration?
Presumably I will need to reboot the ALPHA server as well.
Thanks for the help,
Kind regards, Mark.
I guess so, but remember that a scsi cluster was never possible with a VAX!
The reason for the behaviour is that diskdata and clusterdata follow different paths in a scsi cluster. Which isn't possible on VAXes (I think).
Then again it may be a bug in 8.3 of course!
------Origineel bericht------
Van: Peter Coghlan
Afzender: owner-hecnet at Update.UU.SE
Aan: hecnet at Update.UU.SE
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: Re: [HECnet] Satellite configuration/NCP crash
Verzonden: 16 september 2011 00:37
It should happen before disk access across the CI bus is even possible.
I think we'll have to agree to disagree on this one :-)
I've just booted my Vaxstation 2000 into an NI cluster with my
Vaxstation 3100. I gave the commands:
NCP SET CIRCUIT SVA-0 STATE OFF
NCP SET CIRCUIT SVA-0 STATE ON
on each node. The only effects were an opcom message complaining about
adjacency down and the loss of my decnet connections. Both nodes
continued to function quite happily while the circuit was off. No crash.
Anyway, it looks like avoiding issueing the SET CIRCUIT STATE OFF
command by making the changes to the permanent database and rebooting
was a good way of sidestepping the problem. Hopefully the bug is only
triggered when SET CIRCUIT STATE OFF is issued in Marks setup.
Apologies to gerry77. I missed your reply earlier and ended up repeating
much of what you said.
Regards,
Peter Coghlan.
On 2011-09-16 00:30, Mark Wickens wrote:
Just thinking aloud,
Can I run a bridge between the dongle based network and my own home
network?
If the bridge program on my network is wrapped with a script which
checks the dyndns hostname periodically against the recorded IP address,
cannot I not reconfigure the bridge on my home network and restart it if
required?
If this is possible it keeps my configuration in a realm I've got half a
chance of working out myself and saves me from hassling anyone else.
Presumably at home I would need to run two bridges on different ports -
with the current bridge linking back to Johnny as normal.
Short answer: yes.
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
Mark,
Forget the bridge. Run Multinet at the event and either connect to your
end via Multinet (you are already running that), or connect to mine.
Either way it will just work. I personally, would use the dedicated
events area on the machines at the event and be done with it!
I can either supply you with my code for your end, or you can connect to
my end. Your choice. In the second scenerio, all I have to do is add
one (1) entry in the existing database and off we go!
-Steve
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On
Behalf Of Mark Wickens
Sent: Thursday, September 15, 2011 18:30
To: hecnet at Update.UU.SE
Subject: [HECnet] Another thought regarding dynamic bridge setup
Just thinking aloud,
Can I run a bridge between the dongle based network and my own home
network?
If the bridge program on my network is wrapped with a script which
checks the dyndns hostname periodically against the recorded IP address,
cannot I not reconfigure the bridge on my home network and restart it if
required?
If this is possible it keeps my configuration in a realm I've got half a
chance of working out myself and saves me from hassling anyone else.
Presumably at home I would need to run two bridges on different ports -
with the current bridge linking back to Johnny as normal.
Regards, Mark
It should happen before disk access across the CI bus is even possible.
I think we'll have to agree to disagree on this one :-)
I've just booted my Vaxstation 2000 into an NI cluster with my
Vaxstation 3100. I gave the commands:
NCP SET CIRCUIT SVA-0 STATE OFF
NCP SET CIRCUIT SVA-0 STATE ON
on each node. The only effects were an opcom message complaining about
adjacency down and the loss of my decnet connections. Both nodes
continued to function quite happily while the circuit was off. No crash.
Anyway, it looks like avoiding issueing the SET CIRCUIT STATE OFF
command by making the changes to the permanent database and rebooting
was a good way of sidestepping the problem. Hopefully the bug is only
triggered when SET CIRCUIT STATE OFF is issued in Marks setup.
Apologies to gerry77. I missed your reply earlier and ended up repeating
much of what you said.
Regards,
Peter Coghlan.
Actually, who cares. All of this happens at this end. I can tell this
end to do the lookup every second if I wish. In reality it is usually
done in some number of minutes. As for the reboot, that's why it is a
dedicated machine. It can reboot as often as it needs to - I don't
depend on it except to provide connectivity to this site from remote
sites that have dynamic IP addresses. The actual reboot is only a few
minutes. This entire process is automated and it is able to inform via
email (delayed using NMAIL) anyone necessary that their was a change in
state (address).
-Steve
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On
Behalf Of Johnny Billquist
Sent: Thursday, September 15, 2011 17:46
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] Alternatives to bridge when running via a dynamic
IP address
The question is how long does it take to detect, and is a reboot every
time acceptable?
Johnny
On 2011-09-15 23:40, Steve Davidson wrote:
This issue has been dealt with here. The declab.net site runs a
dedicated VAX with Multinet that polls the other end at predetermined
intervals. When the other ends' address changes this end makes change
to the Multinet database and reboots. This software has been in use
for well over a year and appears to work very well.
It does require the use of DynDNS at the end that changes addresses so
that this end can do the lookup.
-Steve
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On
Behalf Of Mark Wickens
Sent: Thursday, September 15, 2011 17:14
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] Alternatives to bridge when running via a
dynamic IP address
On 15/09/11 22:03, hvlems at zonnet.nl wrote:
Won't the IP address remain unchanged as long as you keep the router
up and running?
Or would thet be too costly?
------Origineel bericht------
Van: Mark Wickens
Afzender: owner-hecnet at Update.UU.SE
Aan: hecnet at Update.UU.SE
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: [HECnet] Alternatives to bridge when running via a dynamic
IP address
Verzonden: 15 september 2011 22:53
Guys,
Might as well explore this as an option.
The 3G network enabled Vigor 2800 router that I will be using at the
DEC Legacy event appears to be working reasonably well, but
unsurprisingly there does appear to be some fluctuation in the IP
address.
This means that an alternative to Johnny's bridge program might be
called for. I have a dynamic DNS reference set up and now updating
periodically, hecnet.no-ip.org, the question is what can I use as an
alternative to the bridge program which will cope with IP address
changes?
I seem to remember Multinet tunnels being mentioned in the past, do
they support dnydns, or any other technology?
Thanks for the help,
Mark.
Unfortunately not. I've seen three IP address changes in about the
same number of hours.
I don't know whether this poor performance is down to the completely
haphazard location of the dongle, or whether it's always this rubbish
but you don't realise when all you're doing is browsing the internet.
My experience of using a 3G dongle for anything other than web
browsing is very limited!
Regards
Mark.
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
Just thinking aloud,
Can I run a bridge between the dongle based network and my own home network?
If the bridge program on my network is wrapped with a script which checks the dyndns hostname periodically against the recorded IP address, cannot I not reconfigure the bridge on my home network and restart it if required?
If this is possible it keeps my configuration in a realm I've got half a chance of working out myself and saves me from hassling anyone else.
Presumably at home I would need to run two bridges on different ports - with the current bridge linking back to Johnny as normal.
Regards, Mark
The question is how long does it take to detect, and is a reboot every time acceptable?
Johnny
On 2011-09-15 23:40, Steve Davidson wrote:
This issue has been dealt with here. The declab.net site runs a
dedicated VAX with Multinet that polls the other end at predetermined
intervals. When the other ends' address changes this end makes change
to the Multinet database and reboots. This software has been in use for
well over a year and appears to work very well.
It does require the use of DynDNS at the end that changes addresses so
that this end can do the lookup.
-Steve
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On
Behalf Of Mark Wickens
Sent: Thursday, September 15, 2011 17:14
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] Alternatives to bridge when running via a dynamic
IP address
On 15/09/11 22:03, hvlems at zonnet.nl wrote:
Won't the IP address remain unchanged as long as you keep the router
up and running?
Or would thet be too costly?
------Origineel bericht------
Van: Mark Wickens
Afzender: owner-hecnet at Update.UU.SE
Aan: hecnet at Update.UU.SE
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: [HECnet] Alternatives to bridge when running via a dynamic
IP address
Verzonden: 15 september 2011 22:53
Guys,
Might as well explore this as an option.
The 3G network enabled Vigor 2800 router that I will be using at the
DEC Legacy event appears to be working reasonably well, but
unsurprisingly there does appear to be some fluctuation in the IP
address.
This means that an alternative to Johnny's bridge program might be
called for. I have a dynamic DNS reference set up and now updating
periodically, hecnet.no-ip.org, the question is what can I use as an
alternative to the bridge program which will cope with IP address
changes?
I seem to remember Multinet tunnels being mentioned in the past, do
they support dnydns, or any other technology?
Thanks for the help,
Mark.
Unfortunately not. I've seen three IP address changes in about the same
number of hours.
I don't know whether this poor performance is down to the completely
haphazard location of the dongle, or whether it's always this rubbish
but you don't realise when all you're doing is browsing the internet.
My experience of using a 3G dongle for anything other than web browsing
is very limited!
Regards
Mark.
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
This issue has been dealt with here. The declab.net site runs a
dedicated VAX with Multinet that polls the other end at predetermined
intervals. When the other ends' address changes this end makes change
to the Multinet database and reboots. This software has been in use for
well over a year and appears to work very well.
It does require the use of DynDNS at the end that changes addresses so
that this end can do the lookup.
-Steve
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On
Behalf Of Mark Wickens
Sent: Thursday, September 15, 2011 17:14
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] Alternatives to bridge when running via a dynamic
IP address
On 15/09/11 22:03, hvlems at zonnet.nl wrote:
Won't the IP address remain unchanged as long as you keep the router
up and running?
Or would thet be too costly?
------Origineel bericht------
Van: Mark Wickens
Afzender: owner-hecnet at Update.UU.SE
Aan: hecnet at Update.UU.SE
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: [HECnet] Alternatives to bridge when running via a dynamic
IP address
Verzonden: 15 september 2011 22:53
Guys,
Might as well explore this as an option.
The 3G network enabled Vigor 2800 router that I will be using at the
DEC Legacy event appears to be working reasonably well, but
unsurprisingly there does appear to be some fluctuation in the IP
address.
This means that an alternative to Johnny's bridge program might be
called for. I have a dynamic DNS reference set up and now updating
periodically, hecnet.no-ip.org, the question is what can I use as an
alternative to the bridge program which will cope with IP address
changes?
I seem to remember Multinet tunnels being mentioned in the past, do
they support dnydns, or any other technology?
Thanks for the help,
Mark.
Unfortunately not. I've seen three IP address changes in about the same
number of hours.
I don't know whether this poor performance is down to the completely
haphazard location of the dongle, or whether it's always this rubbish
but you don't realise when all you're doing is browsing the internet.
My experience of using a 3G dongle for anything other than web browsing
is very limited!
Regards
Mark.
It should happen before disk access across the CI bus is even possible.
------Origineel bericht------
Van: Mark Wickens
Afzender: owner-hecnet at Update.UU.SE
Aan: hecnet at Update.UU.SE
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: Re: [HECnet] Satellite configuration/NCP crash
Verzonden: 15 september 2011 23:29
On 15/09/11 22:27, Peter Coghlan wrote:
Unless youLre running an NI cluster. If PEDRIVER is loaded shutting the decnet circuit used by it must crash the node.
I would have thought that shutting off decnet will not affect SCS
traffic. If it happend that it did, AND there were two or more nodes
in the cluster trying to communicate and failing, I would expect one
or more nodes that lost quorum to hang for RECNXINTERVAL seconds
(by default, 20). After that, one or modes might bugcheck with a
CLUEXIT exception.
A SSRVEXCEPT exception much more likely indicates something is wrong
with the software.
Regards,
Peter Coghlan.
If it helps in the discussion the crash happened very quickly.
