Sender communication channel - Polling delay?

[Former Member]

Hi experts,

We have a sender CC with polling every 6 seconds (other parameters are: connect mode permanently, ftp.timeout 180).

Yesterday a file has been placed on source server at 00.18.00, but CC picked it up only at 00.19.18.

Anyone knows possible reasons for this delay in polling?

I have checked message monitoring there are no errors or connection problems, simply the start and end message date are 00.19.18 and 00.19.19.

Other info:

1) This source server is used also for other interfaces, and in that range of time there is no error, other files have been picked without problems.

2) I'm pretty sure that time of source server is the same respect to XI server, I've tested that.

Thanks

Fabio

[Former Member]

Hi,

I have also faced this isses but failed to find any solid reason for the same.

Once i found a reason to the problem is that when the other system is trying to write the file they used temporary file names which goes from 0% completion to 100% completion and after 100% completion the exact file name is reflected. And XI only pick up the file with the exact match(after 100% completed). So, the XI system try to hit the file multiple times for picking up for processing while the file was written by other system. On multiple hit to the file with the cc, the cc cache refresh problem occurs.

When the cache get refreshed with some other file processing it immediatly try to hit the file and then the file is picked up.

Also, it sometimes happened that the other system tried to write the file with a previous timestamp, the problem might inbetween other system and xilandscape.

Cheers,

Jay

[Former Member]

Hi experts,

any other opinions about this problem?

Fabio

[baskar_gopalakrishnan2]

IMO, if the server time matches with OS then delay should not be there. But there is a possiblity that some other process or thread at oS level might be holding access to that resource (file) during that time due to some valid reasons. Once that access is given back , channel is starting again for polling. You might want to check with system administrators to monitor server log at os level to study the behaviour.

[Former Member]

Hi Fabio,

In the past I was faced to this problem, and my opinion is that polling is however part of a queue, and delay depending from queue's load.

But, i'm waiting with you for a more specific answer...

[Former Member]

Hi Simone,

I agree with you, but in that range of time there were no other files to process in that interface.

There were of course files related to other interfaces polling to the same server and files related to other interfaces related to other servers, do you think that queues are interface-related, server-related or global?

Thanks

Fabio

[Former Member]

I think that is global or server-related, based on Java Instance