In my previous message with regards to journalling I actually meant some records from the journal receivers are not written to a copied Please help
Unconfigured Ad Widget
Collapse
Announcement
Collapse
No announcement yet.
Journalling
Collapse
X
-
Journalling
And how is the file being journaled? INSERTS (output/adds), UPDATES (before and after images) or just afters? and DELETES too? Journaling records for OPEN's and CLOSE's also included? Can you be more specific on what you think is missing...? We're lining up to help you! :-)
Comment
-
Journalling
This is how I started up journaling all the files. STRJRNPF (*libl/File) JRN(*libl/Journal) OMTJRNE(*OPNCLO)...other is default value which is *AFTER image. What I find is the the WRITE record always exist but some of the UPDATE records does not exist. This is causing me some headaches because a main user queried a master record change. Upon interrogating the file that holds the records for the journal receiver I found that many changes done to that master is not in this file although when I check that particular receiver the after change does exist. When I did run a DSPJRN for this particular receiver it does bring in all the records. I hope you are understanding me.
Comment
-
Journalling
[*]AFTER updates are in the JRNRcvr's but NOT in the outfile used when you DSPJRN to the outfile. > As long as you are specifying all of the journal receivers involved, and NOT possibly ommitting any journal entries via parameters of the DSPJRN command...Time to call IBM, I guess? And check for the most current cum PTF's? Sorry I don't have anything for you. Keep us posted on what you find...?
Comment
-
Journalling
In some cases, when a journal entry is written, it is not immediately forced to disk. This happens for operations under commit, and can also occur when you have turned on journal caching. If the journal entry hasn't been forced to disk, DSPJRN won't return them, (since they won't be there if you were to suddenly crash!) To get around this, just before doing your DSPJRN, issue a SNDJRNE FORCE(*YES) to the journal. Also, be sure you are specifying DSPJRN RCVRNG(*CURCHAIN), if you need records older than the current receiver.
Comment
-
Journalling
We are journalling files and all the journalling are done according to my investigation into the journal receivers. Thereafter a CL activated from the jobscheduler every 2 hours runs DSPJRN and which copies the journalled record into a workfile. Thereafter the records from this workfile is added onto another file with CPYF command with *ADD and *NOCHK values and the work files is then deleted. I have a problem that some records in the journal receiver is written to the copied file. Do u think this is the problem with the DSPJRN command.
Code
Comment
-
Journalling
I tried processes similar to what you described and never got past all the timing problems between switching journal receivers, displaying to a file, etc.. I found out that using a program to monitor the journal worked much better. Since my programs are a copy of source developed by someone else, I don't feel free to share the code, but it consisted of an RPGLE program that received records as they came in to the receiver and shutdown commands, a data area that kept track of where to restart from after shutdown, and a constant running program for starting, stopping, etc... The only problem it's had has been if someone shutdown the Monitor and forgets to restart it. It's been a couple of years and I've only done it once. I'll try to come up with some key words unless someone else can point you to some manuals. Good luck
Comment
Comment