Discussion:
HIS 2009 IPDLC stop/start & PU recovery issues
(too old to reply)
stevek
2010-05-27 21:46:54 UTC
Permalink
Testing HIS 2009 IPDLC link in the lab. Ran into situation if you stop then
start IPDLC link, peer and host type PU connections stay INACTIVE/FAILED
under SNA Manager from previous STOP of HIS link. They show CONCT in VTAM.
Have to START each PU connection to get back ACTIVE.

This is different than SNA/SDLC. When you stop then start SNADLC link, the
PU connections automatically recover and go ACITVE by themselves.

PU connections are setup with pretty much same values on the VTAM side. If
there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers? Anyone experience similar
problem?
Stephen Jackson [MSFT]
2010-05-28 16:35:21 UTC
Permalink
Steve,

I just tried this and my connections stayed in a Pending/Failed state for
over an hour. I then came back and started the link service again and the
connections went Active.

Did you change any of the default IP-DLC link service parameters in the link
service properties? Did the connections drop to Inactive/Failed immediately?

Also, as far as I can recall we have never recommended stopping a link
service (DLC or IP-DLC) manually. The link services are setup to start and
stop as the SNA Server service is started and stopped. I'm pretty sure that
we don't do any testing with this type of scenario.

Thanks...
--
Stephen Jackson
Microsoft® HIS Support

Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by stevek
Testing HIS 2009 IPDLC link in the lab. Ran into situation if you stop then
start IPDLC link, peer and host type PU connections stay INACTIVE/FAILED
under SNA Manager from previous STOP of HIS link. They show CONCT in VTAM.
Have to START each PU connection to get back ACTIVE.
This is different than SNA/SDLC. When you stop then start SNADLC link, the
PU connections automatically recover and go ACITVE by themselves.
PU connections are setup with pretty much same values on the VTAM side.
If
there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers? Anyone experience similar
problem?
stevek
2010-05-28 19:13:01 UTC
Permalink
Steve,

We have done similar testing in the past to mimic what whould happen on
failures at the HIS end, including the link. When I tested HIS 2006 SNA DLC
links under the services area, the PU connections always recovered to ACTIVE
status within a couple seconds. I thought I could do similar testing with
the HIS 2009 IPDLC link service and PU connections.

The IPDLC link "PU connections" don't come back until you manually start
them within HIS Manager. The PU connections we use for SNA DLC and IPDLC are
very similar except of course where we have a keyword and value for something
IPDLC.

I remember you pointing out about keeping the default timer values before.
We are using the default IP-DLC lilnk service parameters in the link service
properties: Advanced Tab - IP-DLC timeouts, maximum BTU size, HPR/IP
activation retries.

When the stop the IPDLC link under services, it stops immediately. When I
start it, it starts immediately. SNA Service 1 is not affected by the stops
and starts. IPDLC link service is left as manual as well as manual when we
stop/start SNA DLC link services.

STOP TEST
02:02:20 PM SNA IP-DLC Link Service (event 590) - service terminated by
administrator
02:02:20 PM SNA Server (event 23) - three of these for PU connections (host
or peer). For example: Connection Failure, Connection =PN0BEET0, Link
Service=SNAIP1.

START TEST
02:03:04 PM SNA IP-DLC Link Service (event 626) - service started.
Then there we sit with each of the PU connections under SNA Service 1
showing [Inactive] (Failed @ 14:02:20)

On the host, in LANEET0B we have the main EE PU connection setup. That
responds correctly immediately to me stopping and starting the IPDLC link on
the HIS service. It is the other PU connections in LANEETSK switched major
node that do not come back active. I think I saw something on the web that
says you should keep the IPDLC PU connection separate from the other PU
connections but you can comment on this if it is wrong.

Can you make sure that Microsoft HIS has never had an adverse situation with
the link service failing at the server end. Maybe, we do overkill testing,
but we have always documented this in our HIS test procedures before going
live.

Also, at this point with HIS 2010 beta, it might be worthwhile to switch to
HIS 2010 beta code if we can get it from Microsoft. I will talk to Manager
and the project lead to see if we can switch. Approximately when would HIS
2010 go GA?

It is now 3:12 PM and the PU connections under SNA Manager have not changed.
Will entertain any suggestions you have.
Post by Stephen Jackson [MSFT]
Steve,
I just tried this and my connections stayed in a Pending/Failed state for
over an hour. I then came back and started the link service again and the
connections went Active.
Did you change any of the default IP-DLC link service parameters in the link
service properties? Did the connections drop to Inactive/Failed immediately?
Also, as far as I can recall we have never recommended stopping a link
service (DLC or IP-DLC) manually. The link services are setup to start and
stop as the SNA Server service is started and stopped. I'm pretty sure that
we don't do any testing with this type of scenario.
Thanks...
--
Stephen Jackson
Microsoft® HIS Support
Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by stevek
Testing HIS 2009 IPDLC link in the lab. Ran into situation if you stop then
start IPDLC link, peer and host type PU connections stay INACTIVE/FAILED
under SNA Manager from previous STOP of HIS link. They show CONCT in VTAM.
Have to START each PU connection to get back ACTIVE.
This is different than SNA/SDLC. When you stop then start SNADLC link, the
PU connections automatically recover and go ACITVE by themselves.
PU connections are setup with pretty much same values on the VTAM side.
If
there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers? Anyone experience similar
problem?
Stephen Jackson [MSFT]
2010-05-28 19:48:51 UTC
Permalink
My first test was on a HIS 2006 SP1 (x64) system. I just tried on a HIS 2009
(x86) system. In this case, the "Host" connection dropped to Inactive/Failed
and the Peer connection to Incoming/Failed and then it went to
Pending/Failed.

Once I started the SNAIP1 link service (net start snaip1), both connections
activated automatically within a few seconds.

I'm not sure why you are seeing different behavior.

It might take some tracing to see what it happening in your case.

Thanks...
--
Stephen Jackson
Microsoft® HIS Support

Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by Stephen Jackson [MSFT]
Steve,
We have done similar testing in the past to mimic what whould happen on
failures at the HIS end, including the link. When I tested HIS 2006 SNA DLC
links under the services area, the PU connections always recovered to ACTIVE
status within a couple seconds. I thought I could do similar testing with
the HIS 2009 IPDLC link service and PU connections.
The IPDLC link "PU connections" don't come back until you manually start
them within HIS Manager. The PU connections we use for SNA DLC and IPDLC are
very similar except of course where we have a keyword and value for something
IPDLC.
I remember you pointing out about keeping the default timer values before.
We are using the default IP-DLC lilnk service parameters in the link service
properties: Advanced Tab - IP-DLC timeouts, maximum BTU size, HPR/IP
activation retries.
When the stop the IPDLC link under services, it stops immediately. When I
start it, it starts immediately. SNA Service 1 is not affected by the stops
and starts. IPDLC link service is left as manual as well as manual when we
stop/start SNA DLC link services.
STOP TEST
02:02:20 PM SNA IP-DLC Link Service (event 590) - service terminated by
administrator
02:02:20 PM SNA Server (event 23) - three of these for PU connections (host
or peer). For example: Connection Failure, Connection =PN0BEET0, Link
Service=SNAIP1.
START TEST
02:03:04 PM SNA IP-DLC Link Service (event 626) - service started.
Then there we sit with each of the PU connections under SNA Service 1
On the host, in LANEET0B we have the main EE PU connection setup. That
responds correctly immediately to me stopping and starting the IPDLC link on
the HIS service. It is the other PU connections in LANEETSK switched major
node that do not come back active. I think I saw something on the web that
says you should keep the IPDLC PU connection separate from the other PU
connections but you can comment on this if it is wrong.
Can you make sure that Microsoft HIS has never had an adverse situation with
the link service failing at the server end. Maybe, we do overkill testing,
but we have always documented this in our HIS test procedures before going
live.
Also, at this point with HIS 2010 beta, it might be worthwhile to switch to
HIS 2010 beta code if we can get it from Microsoft. I will talk to Manager
and the project lead to see if we can switch. Approximately when would HIS
2010 go GA?
It is now 3:12 PM and the PU connections under SNA Manager have not changed.
Will entertain any suggestions you have.
Post by Stephen Jackson [MSFT]
Steve,
I just tried this and my connections stayed in a Pending/Failed state for
over an hour. I then came back and started the link service again and the
connections went Active.
Did you change any of the default IP-DLC link service parameters in the link
service properties? Did the connections drop to Inactive/Failed immediately?
Also, as far as I can recall we have never recommended stopping a link
service (DLC or IP-DLC) manually. The link services are setup to start and
stop as the SNA Server service is started and stopped. I'm pretty sure that
we don't do any testing with this type of scenario.
Thanks...
--
Stephen Jackson
Microsoft® HIS Support
Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by stevek
Testing HIS 2009 IPDLC link in the lab. Ran into situation if you stop then
start IPDLC link, peer and host type PU connections stay
INACTIVE/FAILED
under SNA Manager from previous STOP of HIS link. They show CONCT in VTAM.
Have to START each PU connection to get back ACTIVE.
This is different than SNA/SDLC. When you stop then start SNADLC link, the
PU connections automatically recover and go ACITVE by themselves.
PU connections are setup with pretty much same values on the VTAM side.
If
there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers? Anyone experience similar
problem?
stevek
2010-05-28 21:04:01 UTC
Permalink
Steve,

You just provided something that changes the circumstances by mentioning the
Net commands. Instead of stopping/starting SNAIP1 under the Windows 2008 64
bit services area, I am using NET STOP SNAIP1 and NET START SNAIP1 under
command prompt. There are differences.

With all PU connections ACTIVE under SNA Manager, I issue a NET STOP SNAIP1.
SNAIP1 stops and all PU connections go inactive, failed state. When I issue
a NET START SNAIP1, the PU connections that are REMOTE END "HOST" come ACTIVE
now within a few short seconds. You must refresh SNA Manager to get current
state.

However, the one PN0BEET0 PU that is REMOTE PEER based stays inactive,
failed and does not come ACTIVE. There are 3 remote APPC CICS definitions in
HIS server tied to PN0BEET0 but we are not using them for sessions at this
time.

On HIS server, PN0BEET0 is setup as:
Name - PN0BEET0
Link Service - SNAIP1
Comment - anything, doesn't matter
Remote end - Peer system
Allowed directions - Both directions (greyed out)
Activation - On server startup (greyed out)
Affiliate application - None
Support dynamic remote APPC LU defintion - we are not using this right now

We usually don't use NET commands to control link service failure testing
but this has changed everything. I now get consistent recovery on the "host"
PU connections that are connected with SNAIP1 link service.

Should everyting be in one switched major node or two like currently defined
in VTAM? Either way, PN0BEET0 does not come back Incoming or Active.

When your use to SNA SDLC links and HOST PU connections, this is very
interesting
Post by Stephen Jackson [MSFT]
My first test was on a HIS 2006 SP1 (x64) system. I just tried on a HIS 2009
(x86) system. In this case, the "Host" connection dropped to Inactive/Failed
and the Peer connection to Incoming/Failed and then it went to
Pending/Failed.
Once I started the SNAIP1 link service (net start snaip1), both connections
activated automatically within a few seconds.
I'm not sure why you are seeing different behavior.
It might take some tracing to see what it happening in your case.
Thanks...
--
Stephen Jackson
Microsoft® HIS Support
Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by Stephen Jackson [MSFT]
Steve,
We have done similar testing in the past to mimic what whould happen on
failures at the HIS end, including the link. When I tested HIS 2006 SNA DLC
links under the services area, the PU connections always recovered to ACTIVE
status within a couple seconds. I thought I could do similar testing with
the HIS 2009 IPDLC link service and PU connections.
The IPDLC link "PU connections" don't come back until you manually start
them within HIS Manager. The PU connections we use for SNA DLC and IPDLC are
very similar except of course where we have a keyword and value for something
IPDLC.
I remember you pointing out about keeping the default timer values before.
We are using the default IP-DLC lilnk service parameters in the link service
properties: Advanced Tab - IP-DLC timeouts, maximum BTU size, HPR/IP
activation retries.
When the stop the IPDLC link under services, it stops immediately. When I
start it, it starts immediately. SNA Service 1 is not affected by the stops
and starts. IPDLC link service is left as manual as well as manual when we
stop/start SNA DLC link services.
STOP TEST
02:02:20 PM SNA IP-DLC Link Service (event 590) - service terminated by
administrator
02:02:20 PM SNA Server (event 23) - three of these for PU connections (host
or peer). For example: Connection Failure, Connection =PN0BEET0, Link
Service=SNAIP1.
START TEST
02:03:04 PM SNA IP-DLC Link Service (event 626) - service started.
Then there we sit with each of the PU connections under SNA Service 1
On the host, in LANEET0B we have the main EE PU connection setup. That
responds correctly immediately to me stopping and starting the IPDLC link on
the HIS service. It is the other PU connections in LANEETSK switched major
node that do not come back active. I think I saw something on the web that
says you should keep the IPDLC PU connection separate from the other PU
connections but you can comment on this if it is wrong.
Can you make sure that Microsoft HIS has never had an adverse situation with
the link service failing at the server end. Maybe, we do overkill testing,
but we have always documented this in our HIS test procedures before going
live.
Also, at this point with HIS 2010 beta, it might be worthwhile to switch to
HIS 2010 beta code if we can get it from Microsoft. I will talk to Manager
and the project lead to see if we can switch. Approximately when would HIS
2010 go GA?
It is now 3:12 PM and the PU connections under SNA Manager have not changed.
Will entertain any suggestions you have.
Post by Stephen Jackson [MSFT]
Steve,
I just tried this and my connections stayed in a Pending/Failed state for
over an hour. I then came back and started the link service again and the
connections went Active.
Did you change any of the default IP-DLC link service parameters in the link
service properties? Did the connections drop to Inactive/Failed immediately?
Also, as far as I can recall we have never recommended stopping a link
service (DLC or IP-DLC) manually. The link services are setup to start and
stop as the SNA Server service is started and stopped. I'm pretty sure that
we don't do any testing with this type of scenario.
Thanks...
--
Stephen Jackson
Microsoft® HIS Support
Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by stevek
Testing HIS 2009 IPDLC link in the lab. Ran into situation if you stop then
start IPDLC link, peer and host type PU connections stay
INACTIVE/FAILED
under SNA Manager from previous STOP of HIS link. They show CONCT in VTAM.
Have to START each PU connection to get back ACTIVE.
This is different than SNA/SDLC. When you stop then start SNADLC link, the
PU connections automatically recover and go ACITVE by themselves.
PU connections are setup with pretty much same values on the VTAM side.
If
there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers? Anyone experience similar
problem?
Stephen Jackson [MSFT]
2010-05-28 21:37:14 UTC
Permalink
Steve,

I get the same behavior on the when using "Net Stop/Net Start" or when
stopping/starting the service in the Services MMC Snap-in. This was one
three different systems. In each case, the host and peer connections start
automatically once the service is started.

W2K8 (x64)/HIS 2009
W2K8 (x86)/HIS 2009
W2K3 (x64)/HIS 2006 SP1

Strange.

What do you see in the Application Event Log when you stop and start the
service?
--
Stephen Jackson
Microsoft® HIS Support

Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by Stephen Jackson [MSFT]
Steve,
You just provided something that changes the circumstances by mentioning the
Net commands. Instead of stopping/starting SNAIP1 under the Windows 2008 64
bit services area, I am using NET STOP SNAIP1 and NET START SNAIP1 under
command prompt. There are differences.
With all PU connections ACTIVE under SNA Manager, I issue a NET STOP SNAIP1.
SNAIP1 stops and all PU connections go inactive, failed state. When I issue
a NET START SNAIP1, the PU connections that are REMOTE END "HOST" come ACTIVE
now within a few short seconds. You must refresh SNA Manager to get current
state.
However, the one PN0BEET0 PU that is REMOTE PEER based stays inactive,
failed and does not come ACTIVE. There are 3 remote APPC CICS definitions in
HIS server tied to PN0BEET0 but we are not using them for sessions at this
time.
Name - PN0BEET0
Link Service - SNAIP1
Comment - anything, doesn't matter
Remote end - Peer system
Allowed directions - Both directions (greyed out)
Activation - On server startup (greyed out)
Affiliate application - None
Support dynamic remote APPC LU defintion - we are not using this right now
We usually don't use NET commands to control link service failure testing
but this has changed everything. I now get consistent recovery on the "host"
PU connections that are connected with SNAIP1 link service.
Should everyting be in one switched major node or two like currently defined
in VTAM? Either way, PN0BEET0 does not come back Incoming or Active.
When your use to SNA SDLC links and HOST PU connections, this is very
interesting
Post by Stephen Jackson [MSFT]
My first test was on a HIS 2006 SP1 (x64) system. I just tried on a HIS 2009
(x86) system. In this case, the "Host" connection dropped to
Inactive/Failed
and the Peer connection to Incoming/Failed and then it went to
Pending/Failed.
Once I started the SNAIP1 link service (net start snaip1), both connections
activated automatically within a few seconds.
I'm not sure why you are seeing different behavior.
It might take some tracing to see what it happening in your case.
Thanks...
--
Stephen Jackson
Microsoft® HIS Support
Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by Stephen Jackson [MSFT]
Steve,
We have done similar testing in the past to mimic what whould happen on
failures at the HIS end, including the link. When I tested HIS 2006
SNA
DLC
links under the services area, the PU connections always recovered to ACTIVE
status within a couple seconds. I thought I could do similar testing with
the HIS 2009 IPDLC link service and PU connections.
The IPDLC link "PU connections" don't come back until you manually start
them within HIS Manager. The PU connections we use for SNA DLC and
IPDLC
are
very similar except of course where we have a keyword and value for something
IPDLC.
I remember you pointing out about keeping the default timer values before.
We are using the default IP-DLC lilnk service parameters in the link service
properties: Advanced Tab - IP-DLC timeouts, maximum BTU size, HPR/IP
activation retries.
When the stop the IPDLC link under services, it stops immediately.
When I
start it, it starts immediately. SNA Service 1 is not affected by the stops
and starts. IPDLC link service is left as manual as well as manual
when
we
stop/start SNA DLC link services.
STOP TEST
02:02:20 PM SNA IP-DLC Link Service (event 590) - service terminated by
administrator
02:02:20 PM SNA Server (event 23) - three of these for PU connections (host
or peer). For example: Connection Failure, Connection =PN0BEET0, Link
Service=SNAIP1.
START TEST
02:03:04 PM SNA IP-DLC Link Service (event 626) - service started.
Then there we sit with each of the PU connections under SNA Service 1
On the host, in LANEET0B we have the main EE PU connection setup. That
responds correctly immediately to me stopping and starting the IPDLC
link
on
the HIS service. It is the other PU connections in LANEETSK switched major
node that do not come back active. I think I saw something on the web that
says you should keep the IPDLC PU connection separate from the other PU
connections but you can comment on this if it is wrong.
Can you make sure that Microsoft HIS has never had an adverse situation with
the link service failing at the server end. Maybe, we do overkill testing,
but we have always documented this in our HIS test procedures before going
live.
Also, at this point with HIS 2010 beta, it might be worthwhile to
switch
to
HIS 2010 beta code if we can get it from Microsoft. I will talk to Manager
and the project lead to see if we can switch. Approximately when would HIS
2010 go GA?
It is now 3:12 PM and the PU connections under SNA Manager have not changed.
Will entertain any suggestions you have.
Post by Stephen Jackson [MSFT]
Steve,
I just tried this and my connections stayed in a Pending/Failed state for
over an hour. I then came back and started the link service again and the
connections went Active.
Did you change any of the default IP-DLC link service parameters in
the
link
service properties? Did the connections drop to Inactive/Failed immediately?
Also, as far as I can recall we have never recommended stopping a link
service (DLC or IP-DLC) manually. The link services are setup to start and
stop as the SNA Server service is started and stopped. I'm pretty sure that
we don't do any testing with this type of scenario.
Thanks...
--
Stephen Jackson
Microsoft® HIS Support
Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by stevek
Testing HIS 2009 IPDLC link in the lab. Ran into situation if you
stop
then
start IPDLC link, peer and host type PU connections stay
INACTIVE/FAILED
under SNA Manager from previous STOP of HIS link. They show CONCT
in
VTAM.
Have to START each PU connection to get back ACTIVE.
This is different than SNA/SDLC. When you stop then start SNADLC
link,
the
PU connections automatically recover and go ACITVE by themselves.
PU connections are setup with pretty much same values on the VTAM side.
If
there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers? Anyone
experience
similar
problem?
Chris Mason
2010-05-28 21:38:53 UTC
Permalink
Steve

The only possible use for partioning switched PU statements (and associated
PATH and LU statements) between switched major node members is to be able to
(try to) inactivate the resources at the level of the major node. In the dim
and distant past, I seem vaguely to remember needing to do something like
this in order to "force" deactivation. If my memory serves me correctly,
enhancements in the V INACT command long ago removed the need for this
subterfuge.

Thus the recommendation from a long-time VTAM specialist is that this use of
one or two switched major nodes is irrelevant and has no *technical*
significance whatsoever.

Chris Mason
Post by Stephen Jackson [MSFT]
Steve,
You just provided something that changes the circumstances by mentioning the
Net commands. Instead of stopping/starting SNAIP1 under the Windows 2008 64
bit services area, I am using NET STOP SNAIP1 and NET START SNAIP1 under
command prompt. There are differences.
With all PU connections ACTIVE under SNA Manager, I issue a NET STOP SNAIP1.
SNAIP1 stops and all PU connections go inactive, failed state. When I issue
a NET START SNAIP1, the PU connections that are REMOTE END "HOST" come ACTIVE
now within a few short seconds. You must refresh SNA Manager to get current
state.
However, the one PN0BEET0 PU that is REMOTE PEER based stays inactive,
failed and does not come ACTIVE. There are 3 remote APPC CICS definitions in
HIS server tied to PN0BEET0 but we are not using them for sessions at this
time.
Name - PN0BEET0
Link Service - SNAIP1
Comment - anything, doesn't matter
Remote end - Peer system
Allowed directions - Both directions (greyed out)
Activation - On server startup (greyed out)
Affiliate application - None
Support dynamic remote APPC LU defintion - we are not using this right now
We usually don't use NET commands to control link service failure testing
but this has changed everything. I now get consistent recovery on the "host"
PU connections that are connected with SNAIP1 link service.
Should everyting be in one switched major node or two like currently defined
in VTAM? Either way, PN0BEET0 does not come back Incoming or Active.
When your use to SNA SDLC links and HOST PU connections, this is very
interesting
Post by Stephen Jackson [MSFT]
My first test was on a HIS 2006 SP1 (x64) system. I just tried on a HIS 2009
(x86) system. In this case, the "Host" connection dropped to
Inactive/Failed
and the Peer connection to Incoming/Failed and then it went to
Pending/Failed.
Once I started the SNAIP1 link service (net start snaip1), both connections
activated automatically within a few seconds.
I'm not sure why you are seeing different behavior.
It might take some tracing to see what it happening in your case.
Thanks...
--
Stephen Jackson
Microsoft® HIS Support
Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by Stephen Jackson [MSFT]
Steve,
We have done similar testing in the past to mimic what whould happen on
failures at the HIS end, including the link. When I tested HIS 2006
SNA
DLC
links under the services area, the PU connections always recovered to ACTIVE
status within a couple seconds. I thought I could do similar testing with
the HIS 2009 IPDLC link service and PU connections.
The IPDLC link "PU connections" don't come back until you manually start
them within HIS Manager. The PU connections we use for SNA DLC and
IPDLC
are
very similar except of course where we have a keyword and value for something
IPDLC.
I remember you pointing out about keeping the default timer values before.
We are using the default IP-DLC lilnk service parameters in the link service
properties: Advanced Tab - IP-DLC timeouts, maximum BTU size, HPR/IP
activation retries.
When the stop the IPDLC link under services, it stops immediately.
When I
start it, it starts immediately. SNA Service 1 is not affected by the stops
and starts. IPDLC link service is left as manual as well as manual
when
we
stop/start SNA DLC link services.
STOP TEST
02:02:20 PM SNA IP-DLC Link Service (event 590) - service terminated by
administrator
02:02:20 PM SNA Server (event 23) - three of these for PU connections (host
or peer). For example: Connection Failure, Connection =PN0BEET0, Link
Service=SNAIP1.
START TEST
02:03:04 PM SNA IP-DLC Link Service (event 626) - service started.
Then there we sit with each of the PU connections under SNA Service 1
On the host, in LANEET0B we have the main EE PU connection setup. That
responds correctly immediately to me stopping and starting the IPDLC
link
on
the HIS service. It is the other PU connections in LANEETSK switched major
node that do not come back active. I think I saw something on the web that
says you should keep the IPDLC PU connection separate from the other PU
connections but you can comment on this if it is wrong.
Can you make sure that Microsoft HIS has never had an adverse situation with
the link service failing at the server end. Maybe, we do overkill testing,
but we have always documented this in our HIS test procedures before going
live.
Also, at this point with HIS 2010 beta, it might be worthwhile to
switch
to
HIS 2010 beta code if we can get it from Microsoft. I will talk to Manager
and the project lead to see if we can switch. Approximately when would HIS
2010 go GA?
It is now 3:12 PM and the PU connections under SNA Manager have not changed.
Will entertain any suggestions you have.
Post by Stephen Jackson [MSFT]
Steve,
I just tried this and my connections stayed in a Pending/Failed state for
over an hour. I then came back and started the link service again and the
connections went Active.
Did you change any of the default IP-DLC link service parameters in
the
link
service properties? Did the connections drop to Inactive/Failed immediately?
Also, as far as I can recall we have never recommended stopping a link
service (DLC or IP-DLC) manually. The link services are setup to start and
stop as the SNA Server service is started and stopped. I'm pretty sure that
we don't do any testing with this type of scenario.
Thanks...
--
Stephen Jackson
Microsoft® HIS Support
Please do not send e-mail directly to this alias. This alias is for
newsgroup purposes only. This posting is provided "AS IS"
with no warranties, and confers no rights.
Post by stevek
Testing HIS 2009 IPDLC link in the lab. Ran into situation if you
stop
then
start IPDLC link, peer and host type PU connections stay
INACTIVE/FAILED
under SNA Manager from previous STOP of HIS link. They show CONCT
in
VTAM.
Have to START each PU connection to get back ACTIVE.
This is different than SNA/SDLC. When you stop then start SNADLC
link,
the
PU connections automatically recover and go ACITVE by themselves.
PU connections are setup with pretty much same values on the VTAM side.
If
there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers? Anyone
experience
similar
problem?
Chris Mason
2010-05-28 20:24:28 UTC
Permalink
Steve
Ran into situation if you stop then start IPDLC link, peer and host type
PU connections stay INACTIVE/FAILED under SNA Manager from previous STOP
of HIS link. They show CONCT in VTAM. Have to START each PU connection to
get back ACTIVE.
It would appear you have decided that, of the two adjacent link stations,
the VTAM one should be the passive one and HIS should be the active one in
terms of which one accepts connections and which one initiates connections
respectively. This is a normal arrangement where VTAM and a so-called
"distributed" platform are concerned.

However, if it is not possible to have the HIS side recover after a failure,
you could so configure the switched PU statement which represents the VTAM
side of the Enterprise Extender logical link that it takes responsibility to
recover from a failure.

Presumably you are reporting on additional tests with the same configuration
which is the subject of the parallel thread "HIS 2009 SNA Manager not
updating resource status - VTAM Inact For". If you had provided us with your
definitions I could add the necessary statements and operands to your
definitions. As it is I can provide you with a template, "before" and
"after".

"Before"

VBUILD TYPE=MODEL
EEMODEL PU DYNTYPE=EE, Enterprise Extender dynamic link station *
DISCNT=NO, no automatic disconnection *
TGP=name transmission group characteristics

VBUILD TYPE=XCA
name PORT MEDIUM=HPRIP, enterprise extender *
HPREELIV=YES, HPR liveness reduction for EE *
IPRESOLV=0, no resolver timeout *
IPTOS=(20,40,80,C0,C0), EE types of service *
LIVTIME=(20,60), LDLC inoperative timer *
SRQRETRY=4, short request timer retry count *
SRQTIME=3 short request timer interval
grpname GROUP IPADDR=home_address, VIPA for enterprise extender *
AUTOGEN=(n,name,name), maximum connections *
ANSWER=ON, allow incoming connections *
CALL=IN, use for incoming connections *
DIAL=YES, switched procedures required *
DYNPU=YES, dynamic link station control blocks *
KEEPACT=YES reactivate automatically after failure

"After"

VBUILD TYPE=SWNET
name PU CPNAME=name, name of adjacent CP *
DISCNT=NO, no automatic disconnection *
DWACT=YES, connect on activation *
DWINOP=YES, reconnect after INOP *
LIMRES=NO, no automatic deactivation of LU 6.2 *
TGP=name transmission group characteristics
PATH IPADDR=away_address, IP address of adjacent node *
GRPNM=grpname, name of Enterprise Extender XCA GROUP *
CALL=INOUT, use for outgoing and incoming connections *
REDDELAY=60, interval between reconnect retries *
REDIAL=FOREVER retry automatic reconnection

VBUILD TYPE=XCA
name PORT MEDIUM=HPRIP, enterprise extender *
HPREELIV=YES, HPR liveness reduction for EE *
IPRESOLV=0, no resolver timeout *
IPTOS=(20,40,80,C0,C0), EE types of service *
LIVTIME=(20,60), LDLC inoperative timer *
SRQRETRY=4, short request timer retry count *
SRQTIME=3 short request timer interval
grpname GROUP IPADDR=home_address, VIPA for enterprise extender *
AUTOGEN=(n,name,name), maximum connections *
ANSWER=ON, allow incoming connections *
CALL=INOUT, use for outgoing and incoming connections *
DIAL=YES, switched procedures required *
DYNPU=YES, dynamic link station control blocks *
KEEPACT=YES reactivate automatically after failure

where n is equal to or more than the expected number of connections

I have copied the performance-related parameters from a recent project. They
are not always the defaults.
If there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers?
One solution is shown above.

This could solve the "reconnection" problem with the Enterprise Extender
logical link. Once that problem is resolved, you can then check what happens
in the case of the DLUS-DLUR logical link. If this does not recover
automatically, similar changes could be tried out in the definitions for
this logical link. Let us know if that is necessary.
This is different than SNA/SDLC. When you stop then start SNADLC link, the
PU connections automatically recover and go ACITVE by themselves.
As I'm sure the Microsoft HIS specialists will tell you, there is so much
difference between the "IP-DLC Link Service" and earlier types of "Link
Service" that it is not really logical to expect essentially identical
behaviour especially since two different types of logical connection are
involved.
PU connections are setup with pretty much same values on the VTAM side.
Which of the two PU definitions used with the "IP-DLC Link Service" is
"pretty much the same" as with the "DLC Link Service"[1] case? The
Enterprise Extender PU statement - the one I show above - relates to a
vastly different technology. The DLUR PU statement also cannot logically be
compared.

As I have said before, it will help in commenting on your problems when you
show us the definitions you use, possibly indicating any variations you have
tried.

Chris Mason
Testing HIS 2009 IPDLC link in the lab. Ran into situation if you stop then
start IPDLC link, peer and host type PU connections stay INACTIVE/FAILED
under SNA Manager from previous STOP of HIS link. They show CONCT in VTAM.
Have to START each PU connection to get back ACTIVE.
This is different than SNA/SDLC. When you stop then start SNADLC link, the
PU connections automatically recover and go ACITVE by themselves.
PU connections are setup with pretty much same values on the VTAM side.
If
there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers? Anyone experience similar
problem?
stevek
2010-06-03 21:57:58 UTC
Permalink
Chris,

I work on several products so sorry for delays in getting back on some
things. Also, I will be switching to forums/blogs as newsgroups are being
shut down including this HIS one.

In response to your comment about using two separate switched major nodes,
thanks. I'll stick with just one for now.

Have to look at VTAM startup member yet but here are the key values we have
set for APPN EE testing with Microsoft HIS 2009 server.

Host Networking Group has following definitions in place for VTAM
XCACMC1 VBUILD TYPE=XCA
PCMC1 PORT MEDIUM=HPRIP
SAPADDR=04
SRQTIME=15
LIVTIME=15
SRQRETRY=9

Within VTAM Member LANEET0B are the following definitions:
LANEET0B VBUILD TYPE=SWNET
EE0BILAB PU IDBLK=017,IDNUM=EE00A
DYNLU=YES
ADDR=C1
CONNTYPE=APPN
CPCP=YES
ANS=CONT
DLOGMOD=S3278M2
MAXDATA=1033
MODETAB=VTMBIND
SSCPFM=USSSCS
USSTAB=VTAMUSST
PACING=7
MAXOUT=7
PUTYPE=2

TB00ILAB PU IDBLK=017,IDNUM=EE00B
DYNLU=NO
ADDR=C1
ANS=CONT
DLOGMOD=S3278M2
MAXDATA=1033
MODETAB=VTMBIND
SSCPFM=USSSCS
USSTAB=VTAMUSST
PACING=7
MAXOUT=7
PUTYPE=2

Followed by some test LU2, LU6.2 independent, and LU6.2 dependent
definitions.

Per a posting by Steve Jackson with Microsoft, I don't know why I get
differences in PU host/peer recovery when stopping/starting IPDLC link
service under server services area as opposed to using NET START/NET STOP of
the IPDLC link. Will start looking at HIS tracing.

Will share what you presented with our Host Network Group that takes care of
VTAM. Don't know what all the non-defined defaults are yet in the book but
KEEPACT=YES is one that looks interesting.
Post by Chris Mason
Steve
Ran into situation if you stop then start IPDLC link, peer and host type
PU connections stay INACTIVE/FAILED under SNA Manager from previous STOP
of HIS link. They show CONCT in VTAM. Have to START each PU connection to
get back ACTIVE.
It would appear you have decided that, of the two adjacent link stations,
the VTAM one should be the passive one and HIS should be the active one in
terms of which one accepts connections and which one initiates connections
respectively. This is a normal arrangement where VTAM and a so-called
"distributed" platform are concerned.
However, if it is not possible to have the HIS side recover after a failure,
you could so configure the switched PU statement which represents the VTAM
side of the Enterprise Extender logical link that it takes responsibility to
recover from a failure.
Presumably you are reporting on additional tests with the same configuration
which is the subject of the parallel thread "HIS 2009 SNA Manager not
updating resource status - VTAM Inact For". If you had provided us with your
definitions I could add the necessary statements and operands to your
definitions. As it is I can provide you with a template, "before" and
"after".
"Before"
VBUILD TYPE=MODEL
EEMODEL PU DYNTYPE=EE, Enterprise Extender dynamic link station *
DISCNT=NO, no automatic disconnection *
TGP=name transmission group characteristics
VBUILD TYPE=XCA
name PORT MEDIUM=HPRIP, enterprise extender *
HPREELIV=YES, HPR liveness reduction for EE *
IPRESOLV=0, no resolver timeout *
IPTOS=(20,40,80,C0,C0), EE types of service *
LIVTIME=(20,60), LDLC inoperative timer *
SRQRETRY=4, short request timer retry count *
SRQTIME=3 short request timer interval
grpname GROUP IPADDR=home_address, VIPA for enterprise extender *
AUTOGEN=(n,name,name), maximum connections *
ANSWER=ON, allow incoming connections *
CALL=IN, use for incoming connections *
DIAL=YES, switched procedures required *
DYNPU=YES, dynamic link station control blocks *
KEEPACT=YES reactivate automatically after failure
"After"
VBUILD TYPE=SWNET
name PU CPNAME=name, name of adjacent CP *
DISCNT=NO, no automatic disconnection *
DWACT=YES, connect on activation *
DWINOP=YES, reconnect after INOP *
LIMRES=NO, no automatic deactivation of LU 6.2 *
TGP=name transmission group characteristics
PATH IPADDR=away_address, IP address of adjacent node *
GRPNM=grpname, name of Enterprise Extender XCA GROUP *
CALL=INOUT, use for outgoing and incoming connections *
REDDELAY=60, interval between reconnect retries *
REDIAL=FOREVER retry automatic reconnection
VBUILD TYPE=XCA
name PORT MEDIUM=HPRIP, enterprise extender *
HPREELIV=YES, HPR liveness reduction for EE *
IPRESOLV=0, no resolver timeout *
IPTOS=(20,40,80,C0,C0), EE types of service *
LIVTIME=(20,60), LDLC inoperative timer *
SRQRETRY=4, short request timer retry count *
SRQTIME=3 short request timer interval
grpname GROUP IPADDR=home_address, VIPA for enterprise extender *
AUTOGEN=(n,name,name), maximum connections *
ANSWER=ON, allow incoming connections *
CALL=INOUT, use for outgoing and incoming connections *
DIAL=YES, switched procedures required *
DYNPU=YES, dynamic link station control blocks *
KEEPACT=YES reactivate automatically after failure
where n is equal to or more than the expected number of connections
I have copied the performance-related parameters from a recent project. They
are not always the defaults.
If there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers?
One solution is shown above.
This could solve the "reconnection" problem with the Enterprise Extender
logical link. Once that problem is resolved, you can then check what happens
in the case of the DLUS-DLUR logical link. If this does not recover
automatically, similar changes could be tried out in the definitions for
this logical link. Let us know if that is necessary.
This is different than SNA/SDLC. When you stop then start SNADLC link, the
PU connections automatically recover and go ACITVE by themselves.
As I'm sure the Microsoft HIS specialists will tell you, there is so much
difference between the "IP-DLC Link Service" and earlier types of "Link
Service" that it is not really logical to expect essentially identical
behaviour especially since two different types of logical connection are
involved.
PU connections are setup with pretty much same values on the VTAM side.
Which of the two PU definitions used with the "IP-DLC Link Service" is
"pretty much the same" as with the "DLC Link Service"[1] case? The
Enterprise Extender PU statement - the one I show above - relates to a
vastly different technology. The DLUR PU statement also cannot logically be
compared.
As I have said before, it will help in commenting on your problems when you
show us the definitions you use, possibly indicating any variations you have
tried.
Chris Mason
Testing HIS 2009 IPDLC link in the lab. Ran into situation if you stop then
start IPDLC link, peer and host type PU connections stay INACTIVE/FAILED
under SNA Manager from previous STOP of HIS link. They show CONCT in VTAM.
Have to START each PU connection to get back ACTIVE.
This is different than SNA/SDLC. When you stop then start SNADLC link, the
PU connections automatically recover and go ACITVE by themselves.
PU connections are setup with pretty much same values on the VTAM side.
If
there is a IPDLC link failure in HIS, how do you get PU connections to
recover automatically if the IPDLC link recovers? Anyone experience similar
problem?
.
Loading...