Adapter and Channel Monitoring FAQs
1. All Adapters Grayed Out in Adapter Monitor
Q: Since XI 3.0 SP17 / PI 7.0 SP8 most of the adapters and since XI 3.0 SP18 / PI 7.0 SP9 all adapters have been greyed out in the Adapter Monitor. Why did this occur and how can I monitor my adapters in these SPs?
A: Beginning with XI 3.0 SP17 / PI 7.0 SP8 SAP initiated shipping a new monitoring infrastructure for providing enhanced monitoring capabilities for adapters generally and communication channels specifically. Due to which the adapters provided by SAP were all moved over to the new monitoring facility and they don’t supply the old adapter monitoring with status information.
For accessing the new Communication Channel Monitoring, open the Runtime Workbench and select "Component Monitoring" -> "Display" (All) -> "Adapter Engine" -> "Communication Channel Monitor".
2. Channel Monitoring Slow When Retrieving Channel Status
Q: When accessing the Communication Channel Monitoring on my cluster-enabled Adapter Engine functioning on AIX, retrieving the communication channel status takes a very long time to complete and/or states "Error during update". How can I resolve this issue?
A: On an operating system level, set the tunable "tcp_nodelayack" to "1". This can be achieved once the user invokes the command listed as follows: line as "root" user:
no -o tcp_nodelayack=1
3. Channels Keep Old Errors Even After Successful Processing Resumes
Q: For specific sender adapters (like as File, JDBC and Mail) operated in an installation with numerous cluster nodes the aggregated channel status stays "red" ("Processing errors during the last n days") for a long time although successful processing has been resumed in the meantime. The corresponding erratic processing history entries are not expelled. This takes place only intermittently. Why does this occur?
A: Few sender adapters are designed for pining each communication channel to a cluster node. Processing for that channel occurs entirely on this cluster node and only switches to another node if the cluster node fails or the channel configuration has been altered in Integration Builder. The processing history is only kept on each cluster node separately, however, the old entries are not evicted on that node in case the successful processing recommences on a dissimilar node.
4. Manual vs. Automatic Control
Q: Explain the difference between the manual and automatic control?
A: The primary difference between manual and automatic control is that a channel which is under automatic control can be controlled utilizing an availability time (if any exists) whereas the channels under manual control are unaffected by the availability times.