SCACP truncating device IDs

[Brian Reiter]

Hi Folks,

Hopefully this should be a simple vefication. I have a cluster of two RX2800s, running OpenVMS 8-4 and TCPIP 5-7 ECO 3. Each node has 12 network ports (two PCI cards with 4 ports each). The cluster interconnects are via the ports on the motherboard and go to a gigabyte switch. According to TCPIP I have interfaces from IE0 to IE11. Interfaces IE9 and IE10 have been allocated for the cluster traffic (via TCPIP$CONFIG). All seems to be happy and hunky dory until I run SCACP and do a show on the IP devices.

SCACP> sho ip

CATA2 PEA0 Device Summary 13-JUN-2012 12:54:45.09:

         Device   Errors +                               Mgt   Buffer  MgtMax  Line     Total         Current
  Device  Type     Events    Status                   Priority  Size   BufSiz  Speed  Pkts(S+R)    IP  Address
  ------  ----     ------    ------                   --------  ----   ------  -----  ---------     -----------
   IE1                  0  Run Online XChain_Disabled       0   1394       0   1000      267358        10.50.10.82
   IE9                  0  Run Online XChain_Disabled       0   1394       0   1000      267524       10.50.110.82
   IE5                  0  Not in use by SCA                0      0       0    N/A           0            0.0.0.0
   IE6                  0  Not in use by SCA                0      0       0    N/A           0            0.0.0.0
   IE0                  0  Not in use by SCA                0      0       0    N/A           0            0.0.0.0
   IE7                  0  Not in use by SCA                0      0       0    N/A           0            0.0.0.0
   IE8                  0  Not in use by SCA                0      0       0    N/A           0            0.0.0.0

According to SCACP IE1 is in use for the cluster traffic. The details shown match the configuration for IE10. So is this just a truncation problem with SCACP? Will it cause problems elsewhere inside SCACP.

Hopefully should be a simple answer/

cheers

Brian

[John Gillings]

Brian,

   I don't have a system with such an abundance of network adaptors to check your observation, but from what you're saying, it certainly looks like a bug. Perhaps someone has assumed that 10 ethernet devices should be enough for anyone?

  My guess is this is probably mostly cosmetic, the ambiguous names just overwriting each other's displays. What happens if you ask about a specific device?

SCACP> SHOW DEVICE IE10

SCACP> SHOW DEVICE IE1

  If it's a problem, something that might increase your limit to 26 may be to define the devices as single member LAN failover sets (I think that's possible). You would then effectively rename the devices to LLA, LLB, LLC,... (though I'd be wondering, with that many physical devices, why don't you have them all paired up into failover sets anyway? that might also resolve the issue)

Either way, please make sure this is formally reported to OpenVMS engineering.

[Brian Reiter]

Hi John,

It seems to be cosmetic, its just a tad confusing.

SCACP> show ip ie10

CATA1 PEA0 Device Summary 14-JUN-2012 08:21:02.94:

         Device   Errors +                               Mgt   Buffer  MgtMax  Line     Total         Current
  Device  Type     Events    Status                   Priority  Size   BufSiz  Speed  Pkts(S+R)    IP  Address
  ------  ----     ------    ------                   --------  ----   ------  -----  ---------     -----------
   IE1                  0  Run Online XChain_Disabled       0   1394       0   1000      237465        10.50.10.81

I'll pass the issue onto HP. Something may coime of it.

cheers

Brian