Is the Access Database located in local machine or a network shared folder? How many records? From the screenshot, I see there are two components that reading data from Access Database, for debugging, I would suggest to test each step to confirm the issue occur on which component/table? connectivity issue or due to big amount of data？
Could you give us information on access connectors you use ? On the screen shoot we can identify them.
Are all input connectors connected to Access database and only the last output "ABONNEMENT" that is Mysql ?
how is accessed MS Access db ? file access or network connection ?
Can you reproduce the issue with a simple job that you could join to this ticket ?
Did you tried to deactivate all subjob (blue rectangle) and activate again on by one, top to bottom, to find where is the bottleneck ?
What is the size of the lookup ?
What is the link between 'tSetenv' next subjob ? The good way to synchronized subjobs is to use "On subjob OK" between first connector of each two subjob, in your case 'sys_synchro_talend'---- on subjob ok -----> Pink database "ABONNEMENT"
I use tAccessInput with a filter in the query. The Access database is a file on my computer, i don't use a network connection.
In this job, the table "INSTITUT" in the Prejob and the table "ABONNEMENT" with the link "LECTURE" are Access tables. The table "sys_synchro_talend" and the others "ABONNEMENT" tables are MySQL tables.
I tried to deactivate all subjob and activate again one by one but the issue appeared with the first Access component (table "INSTITUT").
So I tried without the table "INSTITUT" in the Prejob. I had no issue with the first subjob or the loading of the lookup "ABONNEMENT", but when the job was connecting to Access table "ABONNEMENT", the CPU usage increased.
For this job, the tables "ABONNEMENT" only have 27 rows (in both database, MySQL and Access).
The link between the component "tSetEnv_2" and the next subjob is a "Run if" link. The next subjob is executed only if the output "CtrlDerSync.Controle"<=0.
I exported the job without the connection information and I joined it to this ticket.
Firs of all I would advise you to declare all db connection in preJob and then use them in all db connectors. If not, "ABONNEMENT" access & mysql connectors will reconnect at each row.
I used the db connection components. The execution is a little faster but i still have the CPU usage issue when Talend is working with Access.
Without the Access connection component, the Access Inputs execution were slow. Now, they are faster but the Access connection component is not (almost 20 seconds). With the MySQL connection component, it's immediate.
For information, I can't see the comments you made in your project because the project linked doesn't have your modifications.
Is it possible for you to reproduce the issue with a simple job ?
[tDBInput(Access)] ---> [tLogRow] ? Try to complexify it until you reproduce.
On the other hand, how many rows is retrieved by the tMap lookup "ABONNEMENT (tDBInput2)(Mysql)" ? Maybe it is the tMap search that increase cpu usage.
Do you really need "All Matches" option for lookup ? Maybe 'First Match' could be ok ?
