netdata control records may have more than one segment

[peter-sylvester]

netdata control records for PDSes created with member names have 2 segments.

The relevant code for control record parsing could be something like

        if 0x20 == (0x20 & flag):
            if 0x80 == (0x80 & flag): 
                crecord_data = b''
                self.logger.debug("[flag 0x80] This is start a control record.")
                record_type = self.xmit_object[loc + 2:loc + 8].decode(self.ebcdic)
                self.logger.debug("Record Type: {}".format(record_type))
            self.logger.debug("[flag 0x20] This is (part of) a control record.")
            crecord_data += self.xmit_object[loc + 2:loc + section_length]  # get the various segments
            if 0x40 == (0x40 & flag):
                self.logger.debug("[flag 0x40] This is end of a control record.")  
                if record_type == "INMR01":  
                    self.parse_INMR01(crecord_data[6:])
                elif record_type == "INMR02":
                    self.parse_INMR02(crecord_data[6:])
                elif record_type == "INMR03":
                    self.parse_INMR03(crecord_data[6:])
                elif record_type == "INMR04":
                    self.parse_INMR04(crecord_data[6:])
                elif record_type == "INMR06":
                    self.logger.debug("[INMR06] Processing last record")
                    return

[wizardofzos]

Aha! So that's what bugging me in the cbt2git code too then maybe?

[peter-sylvester]

Actually, the issue does not seem to create a problem, since member names in the inmr02 are ignored. the actual members for a pds come from the unloaded iebcopy. OTOH, the code should, IMO do better than TSO/E RECEIVE, i.e., when members are present in INMR02, they should match the iebcopy list.
What bug do you experience?