cancel
Showing results for 
Search instead for 
Did you mean: 

SAP CC Bulkloader Processing

mrstifler777
Explorer
678

Hi Experts,

We had an issue regarding SAP CC performance where we there was a huge delay in BITs appearing in CI after rating.

1. We saw in the processing folders (e.g. /workingdirectory/SAPCC_CIT/PREPAID/Instance#1/) that the number of files had build up. Normally the count of files there averages to more or less 250 however in our case it reached 3000+. Is there a way or a guide on how we can determine what's holding up the bulkloader? We think it's waiting for CI but we can't determine what.

2. Also, we have noticed multiple files with bXX (b01) in this processing folder dated more or less a week ago. How many days does CC retain bXX files here? The most recent files doesn't have the bXX anymore. We believe this might have contributed to the delay, and would like to determine why its still in the directory maybe there's an issue. How can we check the logs?

Thanks,

Steve

Accepted Solutions (0)

Answers (1)

Answers (1)

francois_thimon
Product and Topic Expert
Product and Topic Expert

Hi Steve,

Please find some indications below, please let us know if you need more details.

The "bxx" suffix is appended by the bulkloader itself, to indicate that this particular file is currently being processed.
Normally, the suffix therefore stays long enough for the bulkloader to send the file's contents to CI (i.e., usually for a few dozens miliseconds, depending upon your landscape).
A file may keep its "bxx" suffix if something's blocking the bulkloader (e.g., trouble reaching ERP). In that case, the bulkloader may generate thread dumps automatically to describe the problem (and, as you mentioned, each stuck file blocks the queue and delays the subsequent files). The first step is therefore to check the work/dump/ directory for thread dump files whose name contains "timeout", or "exception".
If you know when these new "bxx" files are created, you can also generate thread dumps manually (using "dump bulkLoader#xx" in admin+).
Once you have thread dumps, you may send them to us, and they should tell us what was keeping the bulkloader from completing its process.

You should also check the contents of the files themselves (to confirm that they are valid), and the associated "control files" (i.e., the files that have the same name, but with the ".ctl" extension instead of ".gz" or ".zip" or ".csv").
There control files are used by CC to store detailed error messages, and indications about how much of the item file has already been sent to ERP. They'll at least tell us if the bulkloader had actually started processing the stuck files.

Another thing you may want to check is whether the bulkloader has been rebooted in the past days.
Since the SAP kernel automatically reboots instances, you may not notice such events. The bulkloader is designed to pick up and process old files upon reboot, but an additional issue may have occurred.
So, can you please check the affected bulkloaders' logs too?

As mentioned in the other question here, CC has numerous logs and traces available (JCo, several kinds of thread dumps, but also bulkloading traces). Depending upon the situation, we can assist you in choosing the relevant options to troubleshoot this.
Please let us know what's best.

Best regards.

François
SAP Convergent Charging Support