Offload and service task monitoring

System logger monitors (in 5 second intervals) log-stream-offload activity as well as specific log stream data set allocation requests and notifies the installation if an offload appears to be hung or is taking too long by issuing messages, such as IXG312E, IXG272E, IXG281I, or IXG115A. If this happens, there may be several system logger and non-system logger messages for which responses are expected.

Consider the following steps when responding to these messages:

Determine if there are inhibitors to offload processing by issuing the following commands; then correct any problems:

D LOGGER,C,LSN=logstreamname,DETAIL
D LOGGER,L,LSN=logstreamname
D LOGGER,STATUS,RECALLS
D XCF,STRUCTURE,STRNAME=structname
D GRS,C

If the problem persists, respond to any allocation, catalog, or recall messages such as IEF861I, IEF863I, or IEF458D.
If the DISPLAY LOGGER,STATUS,RECALLS results in message IXG601I revealing that log stream data set recalls are not being satisfied, consider using the SETLOGR FORCE,NORECALL command to cause system logger to stop waiting for a particular data set to be recalled.

If system logger message IXG272E is displayed, reply to this message only after you have attempted to correct any general allocation- or recall- related messages.

If message IXG271I indicates that the delay affects a log stream data set and you reply "FAIL" to the accompanying message IXG272E, log stream duplexing might be affected and the log stream connection might not be able to continue. A reply should only be made when all other avenues have been exhausted.
If the problem persists, respond to any IXG312E messages. This can be used to stop the offload processing for the log stream named in the message and allow it to run on another system, if possible. It might allow other work to run on the system that was attempting the original offload.
If message IXG115A is displayed, reply to this message only after you have attempted to correct any delayed off-loads by responding to the related IXG312E and IXG272E messages. As a last resort, if you reply “TASK=END” to an IXG115A message, system logger ends all the log stream connections in the structure named in the message on this system.

Review the complete description of messages IXG310I, IXG311I, IXG312E, IXG271I, IXG272E, IXG281I, IXG114I, and IXG115A in z/OS MVS System Messages, Vol 10 (IXC-IZP) before responding to any of these messages.

Note that several messages could appear at the same time for different log streams for which offloading is taking place. It may be that only one log stream is actually having a problem, and the others are simply waiting for this log stream to finish its allocation processing. Usually, the log stream that is causing the delay is the first log stream to be reported on.

Another way to determine which log stream is actually causing the delay is to check to see if the data set name ends with an explicit sequence number, such as “A0000001”. If this is the case, it usually means that this log stream is experiencing a problem. If the data set name ends with “.<SEQ#>”, then it is more likely that this log stream is waiting for another log stream that is having a problem.

Start of change Additionally, as interim storage fills, additional messages may be triggered based on the system logger configuration. These messages are intended to monitor interim storage consumption and alert the installation when the capacity reaches critical points. Since offload processing is the method by which interim storage is made available for new data, these are closely related topics. See Monitoring log stream interim storage consumption for further details. End of change

System logger log stream offload activity is monitored using two sets of 2 intervals. The first set of intervals pertain to attempts to allocate or delete log stream offload data sets. The second interval set aids in monitoring migrated data set recall requests during the offload. Within each offload monitor set there is an initial (warning) interval and a secondary (action interval). Logger uses the following default offload monitoring intervals:

For data set allocation/deletion requests, an initial interval of 30 seconds and secondary interval of 60 seconds (1/2 minute and 1 minute, respectively), and
For a recall of a migrated data set request, an initial interval of 60 seconds and secondary interval of 120 seconds (1 minute and 2 minutes, respectively).

When an initial log stream offload monitor interval is reached, system logger will issue message IXG310I to provide an early warning of a delay. When the secondary log stream offload monitor interval is reached, logger issues messages IXG311I and IXG312E to allow possible action on the delayed offload. These two interval values can be specified by the installation. See z/OS MVS Initialization and Tuning Reference for information on providing installation specifications for these log stream offload monitoring intervals in SYS1.PARMLIB member IXGCNFxx.