Rollout/Rollin Processing

When a user program has to be rolled out of thread, Com-plete first attempts to get space in the rollbuffer to simply move the image in storage. If no space or rollbuffer is available, Com-plete attempts to allocate space on the roll dataset(s) to roll the user out to disk. In this case, not enough contiguous space was available on any of the roll datasets.

Com-plete continues processing. The user running the program for which the rollout failed will be informed the next time they cause the Com-plete to try to roll the program in again.

This problem is caused by too many terminals being in use at the same time. Wait a few minutes and try your request again. Contact the Com-plete system programmer about expanding the Com-plete rollout files so that more terminals can be in use at the same time.

Every program on every level requires space to which it can be rolled, either in the rollbuffer or on the roll datasets. This space is only allocated when the program is rolled out and is freed when the program is rolled in again. Refer to the Com-plete System Programmer's Manual for more information on estimates for the roll subsystem.

During rolout processing, the thread for a relocatable program will be prepared so that it can be rolled back into a different thread. If an error occurs during this reloaction of the thread during rollout, this is remembered for the thread. When the thread is rolled back in again for whatever reason or in whatever thread, a check is made to see if the relocation on rolout worked. In this case the relocation has failed.

The program is terminated with a dump and a message is sent to the user using the program.

This is an application or system error which should be reported to the help desk or operations area.

This can happen as a result of an application overwrite in the thread causing Com-plete specific control blocks necessary for relocation to be corrupted. One example of this would be the Com-plete Free Queue Element (FQE) chain. If no evidence of corruption can be found, the problem should be reported to your local support centre, providing the thread dump to assist in the diagnosis of the problem.

A logic error during $2 processing was encountered in the Roll subsystem for tid $1.

The user informed about the error, and where applicable, the program is terminated with a dump.

This is a system error which must be reported to your help desk or operations area.

This basically indicates that Com-plete has got something wrong in the roll subsystem logic. $2 will give an indication as to where this is, however, due to the nature of when the roll subsystem is called, it cannot be certain that the error was due to tid $1, although that terminal will obviously be directly effected by the problem. This should be reported to your local support centre, providing the Com-plete joblog and any dumps produced for diagnostic purposes.

This message is associated with ZRR00002 in that when the application program was last rolled out, the system couldn't find any space on the roll datasets. An event has triggered an attempted rollin of the program, however, as the rollout failed, this message is sent to inform the user.

The user program is terminated and the user is informed.

This error is caused due to the installation defining insufficient resources for Com-plete. Report the problem to you help desk or operations area.

Review the space calculations for the roll subsystem.

In a case where a user program is attempting to start, the previous thread user must first be rolled out of thread so that the new program can run. In this case, the rollout failed and rather than terminate a running session, the inialisation request for the new program is rejected.

The requested program is not started and the requester receives this message to indicate that.

This indicates that there are problems with Com-plete's roll subsystem which will be obvious to the system programmer. Contact your help desk or operations area for more information.

This error will be issued due to a previous rollout error. This can be determined from messages sent in conjunction with this message.

Com-plete has rolled in a copy of a user application program. When the rolling has completed, the tid and level of the program rolled in do not reflect what we expected to be rolled in.

The program is terminated with a dump.

This indicates a system error. Contact your help desk or operations centre.

This is an internal logic error in Com-plete. Collect the Com-plete job log and the dump and contact your local support centre.

Com-plete has rolled a user program back into a thread, however, the number of the thread that is being worked with does not reflect the thread number in relation to what the TIB expected.

The program is terminated with a dump and the user informed.

This indicates a system error. Contact your help desk or operations centre to ensure they are aware of the problem.

This indicates that Com-plete has an internal logic error. Collect the dump and Com-plete joblog and contact your local support centre.

This message is associated with ROL0009 in that the logic error will have occurred while the tib was being rolled out. An event has occurred causing a rolling to be requested and the oppertunity is taken to tell the user about the problem rolling the program out.

The user program is 'terminated' and the user informed.

Please refer to the previously issued ROL0009 as correcting this problem will bypass this message being sent.

This indicates a system error. Contact your help desk or operations area to ensure that they know the problem exists.

When rolling out contents of a thread, Com-plete detected an uncompleted request against the VSAM file indicated by $2.

Action depends on the type of request found active and on whether or not updates to the file indicated are requested to be serialized by Com-plete. 1) If Com-plete serializes updates for the file, and the outstanding request is for update, a snap dump is taken, an ENDREQ issued for the request, and the application is terminated abnormally. 2) In all other cases, processing continues. The message should be treated as a warning about possible deadlock situations.

Generally speaking, applications designed for a multi user environment like Com-plete should not issue terminal I/O or rollout operations while holding any VSAM resources.

Self-explanatory message.