DATALOSS
Advisory
Subject Name Syntax
_RV.ERROR.SYSTEM.DATALOSS.OUTBOUND.PTP
_RV.ERROR.SYSTEM.DATALOSS.OUTBOUND.BCAST
_RV.ERROR.SYSTEM.DATALOSS.INBOUND.PTP
_RV.ERROR.SYSTEM.DATALOSS.INBOUND.BCAST
Purpose
Client programs receive this advisory when a sending daemon denies a retransmission request.
Remarks
DATALOSS
advisories indicate network transmission problems.
Rendezvous daemons use a reliable delivery protocol, in which sending daemons retain outbound messages for a limited time (called the reliability interval or message retention time). If a receiving daemon detects that it missed an inbound message, it requests retransmission from the sending daemon. If the retention time has already elapsed, the sending daemon has already discarded the message, so it cannot retransmit. Under normal operating conditions, both daemons notify all their client transports that data has been lost; in some situations, not all of the daemons report the loss.
Clients of the sending daemon present DATALOSS.OUTBOUND
advisories. Clients of the receiving daemon present DATALOSS.INBOUND
advisories.
PTP
indicates that the lost data was a point-to-point message. BCAST
indicates that the lost data was a multicast message.
Since the daemons cannot determine which client transports are affected by the loss, they present these advisories to all of their clients on the service that lost data (even though not every client on that service has actually lost data).
(The factory default retention time is 60 seconds. For a complete discussion the concept of reliability, the various ways to control it, the interaction among those ways, and reasonable values, see Reliability and Message Retention Time on page 29 in .)
Message Fields
Field Name |
Description |
|
The IP address of the other computer. |
|
The number of packets requested by the host, but not retransmitted by the sending daemon (during the interval since the last advisory of this type for the receiving host and service). |
|
Service communication ID. For internal use only. This value has type |
Diagnosis
These advisories indicate situations that defeat the Rendezvous reliable delivery protocols:
• | Some hardware component is experiencing intermittent failures; the component could be a faulty network card, a loose connection, or a frayed wire. |
• | The network is overloaded. |
• | A Rendezvous daemon process is starved for CPU cycles; either its host computer is too heavily loaded, or the priority of the daemon process is too low. |
• | The Rendezvous daemon is running with a reliability interval that is too low. (See also, Reliability and Message Retention Time in .) |
Limitations
Rendezvous reliable delivery protocols implement fast and efficient delivery of messages under normal operating conditions. For diagnostic convenience, the Rendezvous daemon reports DATALOSS
advisories when detection would not incur the cost of additional network traffic. However, DATALOSS
advisories are not guaranteed in every situation.
Rendezvous software does not report DATALOSS
advisories across routing daemon neighbor links—only to transports directly connected to the daemon that detects the loss.
If your program requires stronger confirmation of delivery, consider using the certified delivery feature (see Certified Message Delivery).