Explanation
This unsolicited message is issued when HPR detects
a transmission stall for an RTP pipe. A transmission stall occurs
when the RTP partner continuously requests transmission of the same
network layer packet (NLP). VTAM® honors
the request each time by transmitting the NLP, but the partner does
not receive it. This message is issued after VTAM retransmits the NLP for the sixth time
and at least 4-10 seconds has elapsed since the NLP was first retransmitted.
In the message text:
- puname
- The name of the HPR PU.
- cpname
- The name of the CP at the other end of the pipe.
System action
When a specific NLP is retransmitted 6 times
without being acknowledged by the RTP partner, and at least 10 seconds
has elapsed since the NLP was first retransmitted, VTAM notifies the operator with this message.
The system actions vary depending on whether the RTP pipe traverses
an Enterprise Extender (EE) link.
- If the RTP pipe traverses EE, VTAM provides
RTP transmission stall recovery support. If the RTP connection is
multiple hops with at least one refifo connection, an HPR path switch
is started. If the RTP connection is one hop over EE or an EE connection
network, the RTP endpoint immediately lowers its NLP size to the minimum
768 bytes allowed by the HPR architecture. All outbound data is then
re-segmented to this value to resolve the transmission stall. If the
stall persists, VTAM issues
message IST2246I every 30 seconds until the stall is alleviated. If
a path switch was initiated to alleviate the stall and the stall persists, VTAM lowers the NLP size to 768
bytes when message IST2246I is issued. If the stall is alleviated, VTAM issues message IST2247I and
normal data flow resumes. After 20 minutes of operating with the reduced
NLP size, the original NLP size of the RTP pipe is restored to increase
RTP throughput. If the network still does not allow transport of the
larger packets, the stall detection and alleviation process repeats.
- If the RTP pipe does not traverse EE, VTAM does not provide RTP transmission stall
recovery support. In these configurations, the maximum NLP size is
calculated for the path when the RTP pipe is established. The underlying
NLP size for the RTP path should not change during the life of the
RTP pipe, unless a path switch occurs. VTAM alerts
the operator of a transmission stall and continues to honor the retransmission
requests sent by the RTP partner. If the stall persists, VTAM issues message IST2246I every
30 seconds until the stall is alleviated. If the stall is alleviated, VTAM issues message IST2247I and
normal data flow resumes.
If the stall persists beyond the time limit specified
by the HPRSTALL start option, VTAM automatically
deactivates the RTP pipe. See the information about the HPRSTALL start option in z/OS Communications Server: SNA Resource Definition
Reference.
Operator response
Perform the following steps to alleviate
the transmission stall:
- If the RTP pipe traverses EE and the following conditions are
true, verify that the firewalls are configured to allow ICMP errors
to flow on all hops of the connection:
If the firewalls are not allowing ICMP errors to flow, this
could be the cause of the transmission stall. Disable path MTU discovery
for EE connections until the firewalls are configured to allow ICMP
errors to flow:
- Issue the MODIFY procname,VTAMOPTS,
PMTUD=NO command to temporarily disable path MTU discovery.
- Contact the system programmer to permanently disable path MTU
discovery for IPv4 and IPv6 EE connections.
- If the transmission stall persists, consider ending the stalled
sessions by issuing the VARY net,INACT,ID=rtp_pu_name,TYPE=FORCE command
to deactivate the HPR PU. If you want VTAM to
automatically end stalled RTP pipes after a specified amount of time,
use the HPRSTALL start option. See the HPRSTALL
start option information in z/OS Communications Server: SNA Resource Definition
Reference.
System programmer response
If you want path MTU discovery for
IPv4 and IPv6 Enterprise Extender connections to be permanently disabled,
specify PMTUD=NO in the appropriate ATCSTRxx VTAM start list. If you want path
MTU discovery for IPv4 and IPv6 Enterprise Extender connections to
be enabled, ensure that all firewalls in the network path are configured
to allow ICMP errors to flow.
User response
Problem determination
If transmission stalls are prevalent
and persistent, contact VTAM support
for instructions about gathering documentation for problem determination.
- If you have access to IBMLink, search for known problems with similar
symptoms.
- If no applicable matches are found, or if you do not have access
to IBMLink,
obtain a diagnostic dump for this error and report the problem to IBM®. If you have access to IBMLink, the
problem can be reported to IBM using
the Electronic Technical Report (ETR) option on IBMLink.
Source
z/OS® Communications
Server SNA
Module
Routing code
Descriptor code
Automation
Example
IST2245I XMIT STALL DETECTED FOR RTP CNR0002F TO NETA.SSCP1A