ERecruit 604 slowness

Question

Hi,
We have upgraded our eRecruit from 600 to 604 and since after the upgrade, the users are experiencing extreme slowness in certain areas of the system especially "job posting". When a recruiter tries to post a job or search for a candidate, the system gets extremely slow. As I observed this happens when the HRRCFCAND and HRRCFCDCY jobs are running for longer durations on the system. Though these jobs generally run not more than 15 minutes (Our Periodical Services is scheduled every 15mins), there are some instances during the day, where they run for more than 45 mins or 60 mins. At these times, the users exprience extreme slowness where every click can take upto 4-5 mins.
The reason as I've observed is, the jobs are updating the INDX table and even the users are trying to update at the same time. I guess eRec is designed in such a way that every click of the user has to first update INDX table. That is suprising, even when the users are still browsing just to get to the point where they can start posting, they are already updating INDX at the backend. Now, as I said when that coincides with the HRRCF jobs updating the same table, system is darn slow. We've opened a message with SAP and they suggested us to apply note 1091587. We've also applied notes 1325415, 1316416 and 1338545 depending on the scenarios. But nothing seems to work.
Would anybody know why this happens? Is this a bug in 604 SP2?
Appreciate your responses.
-Vishnu

Answer 1

Hi Vishnu,
well it is not really a bug, but a bad side effect of a BAPI which is used within e-recruiting.
If you index candidates, the root node of the qualification catalogue is retrieved for each candidate and saved in the INDX cluster. If you index many candidates the indexing has an exclusive lock of the cluster. Every candidate who wants then see the PDF data overview, has to wait until indexing is over. The note should help because it keeps the catalogue node in a buffer while indexing.
If the note does not help, you can customize the  "restrict view on the qualification catalogue" in the IMG. This will help definetely, because then the root node is not saved in the cluster anymore. In this customizing view you need to type in every 1st level node of the qualification catalogue, you want to make visible to the candidate. But be careful. Just do it, if the note does not help and you will have to do it on each system, because QK IDs are different on each system!
Best regards
Sebastian