Crit messages unsupported version 0x0 != 0x1 from transport/strmsock.cpp

[openhpi2]

Hi Anton,

Greetings..

We are using OpenHPI version 3.2.0

Below critical logs are observed during restart of management-blade, which are very rare occurrence.

Sep 29 12:00:32.231478 crit CLA-0 openhpid: transport: /var/fpwork/fullbuild-jenkins/workspace/R_FPT_41.4.1.x-rc/variant/wr-x86_64-j/SS_OpenHPI/openhpi/transport/strmsock.cpp:198: unsupported version 0x0 != 0x1.

Sep 29 12:00:46.826558 crit CLA-0 openhpid: transport: /var/fpwork/fullbuild-jenkins/workspace/R_FPT_41.4.1.x-rc/variant/wr-x86_64-j/SS_OpenHPI/openhpi/transport/strmsock.cpp:198: unsupported version 0x0 != 0x1.

From code, could see comparison is made b/n ver & dMhRpcVersion variables.

      uint8_t ver = hdr[dMhOffFlags] >> 4;
      if ( ver != dMhRpcVersion ) {
            CRIT( "unsupported version 0x%x != 0x%x.",
                 ver,
                 dMhRpcVersion );
            return false;
        }

Could you please throw some light on various possible situations, where above messages would be displayed.

Best Regards,
Mahesh.

Reported by: nj1213

[openhpi2]

Mahesh,
I do not know if this is a bug. Could you please verify is this critical message still exists in 3.6.1?
Thanks
Mohan

Original comment by: dr_mohan

[openhpi2]

I cannot imagine any sane situation for the issue to happen.
See transport/strmsock.cpp, functions ReadMsg and WriteMsg.
And the version constant was never changed on my memory for about 10 years.

If remote party drops connection, then there is "error while reading message" log. Do you see it as well?

Btw, what is being restarted - openhpid system or HPI client app system?

Original comment by: avpak

[openhpi2]

Original comment by: dr_mohan

[openhpi2]

Mahesh, do you see this problem in 3.6.1 also? Could this be closed?

Original comment by: dr_mohan

[openhpi2]

Anton, is it ok to close this issue?

Original comment by: dr_mohan