TSM scheduler using the wrong IP address for a node

Hi there,

we're using TSM 5.2.2.2 on AIX 5.2. The client is at 5.2.2.5, also on AIX
5.2. (Both 64 bit)

I cloned a machine using mksysb. On the clone, I changed the IP address.
I accidentally started the dsmcad process with an unchanged configuration.
This caused the TSM server to update the information it has for that node.
This was shown in a "q nod tridev1 f=d". I stopped and disabled the TSM
installation on the clone, and did a manual backup with the original
machine.

This worked fine, and both 'q nod tridev1 f=d' and "select * from nodes
where node_name='TRIDEV1'" now show the correct IP address. However: the
scheduler still insist on using the wrong IP address. Given the output of
"select * from nodes where node_name='TRIDEV1'" below, how is this
possible?

         NODE_NAME: TRIDEV1
     PLATFORM_NAME: AIX
       DOMAIN_NAME: PD_TRIAS
        PWSET_TIME: 2003-10-29 14:27:10.000000
  INVALID_PW_COUNT: 0
           CONTACT:
       COMPRESSION: CLIENT
        ARCHDELETE: YES
        BACKDELETE: NO
            LOCKED: NO
      LASTACC_TIME: 2004-04-23 09:22:21.000000
          REG_TIME: 2001-06-11 13:43:07.000000
         REG_ADMIN: MM
 LASTSESS_COMMMETH: Tcp/Ip
    LASTSESS_RECVD: 15210894
     LASTSESS_SENT: 17493
 LASTSESS_DURATION: +3.96090000000000E+001
 LASTSESS_IDLEWAIT: +1.43906687873968E-001
 LASTSESS_COMMWAIT: +4.57724254588603E+000
LASTSESS_MEDIAWAIT: +0.00000000000000E+000
    CLIENT_VERSION: 5
    CLIENT_RELEASE: 2
      CLIENT_LEVEL: 2
   CLIENT_SUBLEVEL: 5
   CLIENT_OS_LEVEL: 5.2
        OPTION_SET:
       AGGREGATION: YES
               URL: http://tridev1.trias.nl:1581/
          NODETYPE: CLIENT
           PASSEXP: 9999
           KEEP_MP: NO
    MAX_MP_ALLOWED: 3
    AUTO_FS_RENAME: NO
  VALIDATEPROTOCOL: NO
          TCP_NAME: tridev1
       TCP_ADDRESS: 172.17.1.42
              GUID: 05.50.1b.d0.d8.9e.11.d7.9b.74.08.63.ac.11.01.8b
       TXNGROUPMAX: 0
     DATAWRITEPATH: ANY
      DATAREADPATH: ANY
SESSION_INITIATION: ClientOrServer
        CLIENT_HLA:
        CLIENT_LLA:


Only after executing "UPD NOD TRIDEV1 HLADDRESS='172.17.1.42' LLADDRESS='1500'",
the scheduler used the correct address.

There's also something wrong with UPDATE NODE: if you issue "UPD NOD
TRIDEV1 HLADDRESS='172.17.1.41'" the response is "Node updated", but that
is in fact not the case. In the actlog an error is reported:

04/23/04 10:39:07     ANR2017I Administrator JOSKAM issued command: UPDATE NODE
                        tridev1 hladdress=172.17.1.42  (SESSION: 26308)
04/23/04 10:39:07     ANR4739E Either the high level address ( 172.17.1.42 ), l
                       ow level address( ? ), or both were entered in error.  (
                       SESSION: 26308)
04/23/04 10:39:07     ANR2063I Node TRIDEV1 updated. (SESSION: 26308)

You don't get to see the ANR4739E unless you look in the actlog.

--
Jurjen Oskam

"Avoid putting a paging file on a fault-tolerant drive, such as a mirrored
volume or a RAID-5 volume. Paging files do not need fault-tolerance." - 308417