DSXchange

Hi,
I am getting this error.
.INPUT_TO_AGGR: ds_ipcput() - timeout waiting for mutex
I really appreciate if someone telles me how to solve this problem.

thanks
shivan

shivan wrote:Hi,
I am getting this error.
.INPUT_TO_AGGR: ds_ipcput() - timeout waiting for mutex
I really appreciate if someone telles me how to solve this problem.

thanks
shivan

Hi shivan,
tell more about the issue, excatly where u get this error, or what r trying to do.
I could understand the error is to the aggregator stage,
but..............

regards
kumar

Hi,
thanks for replying. I am trying to get the data from db2 and putting into mainframe. The number of records are more than 1 million. It works fine while going from db2 to transformer and then to aggregate processor. But at the time of writing to mainframe. It gives an error. I was thinking of something has to do with array size which is 1 right now. Or transaction handling which i set it up to 250.
I hope this helps u in understanding my problem
thanks
shivan

kumar_s wrote:

shivan wrote:Hi,
I am getting this error.
.INPUT_TO_AGGR: ds_ipcput() - timeout waiting for mutex
I really appreciate if someone telles me how to solve this problem.

thanks
shivan

Hi shivan,
tell more about the issue, excatly where u get this error, or what r trying to do.
I could understand the error is to the aggregator stage,
but..............

regards
kumar

Disable inter process comm and retry.

Inter process is already disabled. Settings is to default.

thanks
shivan

Sainath.Srinivasan wrote:Disable inter process comm and retry.

I run the job again. Now i got this error:
CLI0108E Communication link failure. SQLSTATE=40003

thanks

shivan wrote:Inter process is already disabled. Settings is to default.

thanks
shivan

Sainath.Srinivasan wrote:Disable inter process comm and retry.

hi shivan,

Looks like you connection timed out, I believe your source is returning
more data and it takes time read all the rows from the source by the
time the target is disconnected because of it reached the idle time.
Please pick up fewer amounts of data and test them and talk to your DBA
about the connection idle time.

regards
kumar

thanks Kumar, i will check with a DBA. the other problem i have is performance , the number of rows are like 800 per sec. Is there any way i can make them more.

thanks
shivan

kumar_s wrote:hi shivan,

Looks like you connection timed out, I believe your source is returning
more data and it takes time read all the rows from the source by the
time the target is disconnected because of it reached the idle time.
Please pick up fewer amounts of data and test them and talk to your DBA
about the connection idle time.

regards
kumar

Hi Shivan,
Use a SORT stage before the aggregator and sort the columns. Also make sure you mention the sort order in the Aggregator.

If you can split the query into smaller time frames thats great, else try to extract into a flat file and then load it in another job, which is not a great idea as you will be landing the data which is not required but it would be easy for you to test it , instead of querying it everytime it fails!

hi,
i am getting this error :
Abnormal termination of stage CopyOfDB2_INVOICE_LINE_ITEM_RPT_part..Aggregator3 detected
I am using link_partitioner and partioner_collector.

thanks
shivan

vinaymanchinila wrote:Hi Shivan,
Use a SORT stage before the aggregator and sort the columns. Also make sure you mention the sort order in the Aggregator.

If you can split the query into smaller time frames thats great, else try to extract into a flat file and then load it in another job, which is not a great idea as you will be landing the data which is not required but it would be easy for you to test it , instead of querying it everytime it fails!

Hi,
You mean you are using them before and after the aggregator?
And are you sorting the data before passing to aggreVator!

shivan wrote:hi,
i am getting this error :
Abnormal termination of stage CopyOfDB2_INVOICE_LINE_ITEM_RPT_part..Aggregator3 detected
I am using link_partitioner and partioner_collector.

thanks
shivan

vinaymanchinila wrote:Hi Shivan,
Use a SORT stage before the aggregator and sort the columns. Also make sure you mention the sort order in the Aggregator.

If you can split the query into smaller time frames thats great, else try to extract into a flat file and then load it in another job, which is not a great idea as you will be landing the data which is not required but it would be easy for you to test it , instead of querying it everytime it fails!

no, i am not. If this is what i m doing wrong. I will do that. I thought that sort is to increase the performance.

thanks
shivan

vinaymanchinila wrote:Hi,
You mean you are using them before and after the aggregator?
And are you sorting the data before passing to aggreVator!

shivan wrote:hi,
i am getting this error :
Abnormal termination of stage CopyOfDB2_INVOICE_LINE_ITEM_RPT_part..Aggregator3 detected
I am using link_partitioner and partioner_collector.

thanks
shivan

vinaymanchinila wrote:Hi Shivan,
Use a SORT stage before the aggregator and sort the columns. Also make sure you mention the sort order in the Aggregator.

If you can split the query into smaller time frames thats great, else try to extract into a flat file and then load it in another job, which is not a great idea as you will be landing the data which is not required but it would be easy for you to test it , instead of querying it everytime it fails!

hi,
i m not using the sorting at all. My job looks like this:
db2 then a partitioner then three transformers and then three aggregators and then link collector and then mainframe database. Now i m getting this error:
ds_ipcgetnext - timeout waiting for mutex
There are like 1.7 million records. Can i run three jobs based on different dates, which divides the data. I dont think so it is acceptable. But i am not sure.
Your help is really appreciated.
thanks
shivan

shivan wrote:no, i am not. If this is what i m doing wrong. I will do that. I thought that sort is to increase the performance.

thanks
shivan

vinaymanchinila wrote:Hi,
You mean you are using them before and after the aggregator?
And are you sorting the data before passing to aggreVator!

shivan wrote:hi,
i am getting this error :
Abnormal termination of stage CopyOfDB2_INVOICE_LINE_ITEM_RPT_part..Aggregator3 detected
I am using link_partitioner and partioner_collector.

thanks
shivan

Hi Shivan,

I m not using the sorting at all.

Passing unsorted input data to an Aggregator, might work for a small number of rows, as Aggregator internally sorts it and then applies the aggregate functions to the data.

However, it might not work in case of a large number of rows. Hence it is highly recommended to sort the data in a sort stage and also make sure you mention the same sort key in the Aggregator stage.

There are like 1.7 million records. Can i run three jobs based on different dates, which divides the data.

Yes, you can run three jobs with the input to each job being a subset of the complete data, obtained by your date criteria.

Gurus, pls correct me if I am wrong in any of the above comments.

I had the same issue with the aggregator when I did NOT use the Sort stage, and I had 2.5 million records.

I sorted the data and it now works fine!