Performance problem with R4I on AS400

[giocot]

Hi all

i would need help from  some one that already and installed and used R4I for understand how you have configured in AS400 R4I (Subsystem, memory ect) as when  we use R4I , we have performance problem (high Cpu usage) .

can some help us to solve the problem ?

Gio

[john_wang]

Hello Gio, @giocot 

We need additional information to better understand the behavior:

1. How many tasks are running in parallel and connecting to the same AS/400 system?
2. How many tables are included in each task?
3. Were the tables added using “Table Selection Patterns”?
4. What is the journal volume size during the high CPU usage period?

BTW, CPU usage on AS/400 is not directly comparable to CPU usage on Linux or other operating systems.

For example, on Linux, CPU% generally reflects the total CPU cycles consumed across all cores. On AS/400, however, CPU% may represent usage within a specific shared pool or subsystem. Depending on how jobs and threads are accounted for, the reported value can even exceed 100%.

Regards,

John.

[giocot]

Hi John 

thanks for you reply ..  be patient (i don't know Qlik..) so i would ask to better explain this point:

"Were the tables added using “Table Selection Patterns”? "  i don't well understand 

" What is the journal volume size during the high CPU usage period?"  how can get this information ?

Thanks in advance

Gio 

P.S. just for information: which logic is used by the UDTF r4I, to read journal record ? i try to better explain:in journal there are varoius record for any table changes, when a record i read from R4I, how do to skip (the record already read) next time that R4I exceute the selection ??

[john_wang]

Hello Gio, @giocot 

Were the tables added using “Table Selection Patterns”?

Tables can be added either explicitly or through selection patterns, for example:

In the above example, all tables beginning with KIT* as well as the table TESTEMPTY will be included.

What is the journal volume size during the high CPU usage period?

Please work with your DBA to collect this information. It can usually be checked either from the IBM i terminal or through IBM i Navigator.

Which logic is used by the UDTF r4I to read journal records?

Qlik Replicate uses the DISPLAY_JOURNAL() API to capture changes from the AS/400 journal.

how do to skip (the record already read) next time that R4I execute the selection ??

Qlik Replicate maintains the stream position of previously processed journal records in both its runtime memory and repository, allowing subsequent reads to continue from the correct position without rereading processed records.

Thanks,

John.