When trying to use the z/OSMF console interface to issue a console command and get the response I got after 60 seconds.
{“reason”:”Timeout when creating TSO address space for console COLIN99″,”return-code”:2,”reason-code”:21}
This was caused by PARMLIB(IZUPRMCM) not having a HOSTNAME specified, and was picking up a HOSTNAME of S0W1.DAL-EBIS.IHOST.COM. When I specified HOSTNAME(10.1.1.2) and restarted z/OSMF it all worked!
The HOSTNAME can be in various places in TCP/IP and the value from TSO (and jobs) may be different from a Unix thread.
I had to use the address of the TCP/IP stack for example 10.1.1.2; using the address of 127.0.0.1 didn’t work. I could not connect to z/OSMF from my client.
It was strange that with the HOSTNAME missing, I could create TSO address spaces.
How did I diagnose this problem?
In the file /global/zosmf/data/logs/IZUG0.log were entries like
INFO:Prepare to start new TSO/E address space with acct: ACCT#, proc: IZUFPROC, rsize: 50000, apptag: IZUCONAP
Ýtx000000000000000E:IBMUSER@10.1.0.2 (PUT) /zosmf/restconsoles/consoles/COLIN99?null¨
2025-06-28T16:44:02.114Z|00000092|com.ibm.zoszmf.consoles.tsoconnect.Connection|run
WARNING:exception when run as server:
java.net.SocketTimeoutException: Connect timed out
...
at java.base/sun.net.www.http.HttpClient.openServer(HttpClient.java:531)
...
at com.ibm.zoszmf.consoles.tsoconnect.Connection$3.run(Connection.java:372)
This shows the TSO account number, the TSO procedure, the region size. I checked these were valid, and the userid had access to them.
The http request was a clue that this was using TCP/IP, and not using the TSO services available through CEA interface.
How did I run the command?
I used a script like
name=”colinpaice”
cert=” –cert ./$name.pem:password –key $name.key.pem”
insecure=”–insecure”
curl -X PUT –header ‘Content-Type: application/json’ $cert –header ‘Accept: application/json’
–insecure -d ‘{ “cmd”: “d a,l”, “sol-key”: “JES” }’
‘https://10.1.1.2:10443/zosmf/restconsoles/consoles/COLIN99’
ColinPaice 