[omniORB] Sense of reinvoking the same method

[Ridgway, Richard (London)]

This depends where you expect the delay to occur. If the invocation
always takes < 10-40ms to process in the remote object then you are ok.
If this could take longer when under load then retrying in the way you
suggested is only likely to make the situation worse by adding yet more
load to the remote server - you would be better off allowing the maximum
500ms for the call to return.
If on the other hand you are trying to deal with network problems, it
sounds reasonable - but bare in mind that omniorb performs retries
automagically under the covers in situations it recognises.


Richard


-----Original Message-----
From: omniorb-list-bounces at omniorb-support.com
[mailto:omniorb-list-bounces at omniorb-support.com] On Behalf Of
radamkie at kdm.pl
Sent: 29 January 2007 13:11
To: omniorb-list at omniorb-support.com
Subject: [omniORB] Sense of reinvoking the same method


Hi,

I would like to ask about a situation when invoking a remote method
takes much longer time that I need. There is a way to "cancel" such
method by setting timeout. But I have a question, if there is still a
sense
to try re-invoking method. For example:

I need to invoke remote method functionX() in 500 ms, which normally
takes
10-40 ms. So I plan to do this in such way:

- invoke method functionX() with timeout set to 200 ms
- if invoking takes longer time then 200 ms then
"cancel" (by timer) and invoking the same method second time
and count that this time it takes less time then 300 ms.

Has it sense ? Mayby somebody have experience in it ?

Regards
R.






_______________________________________________
omniORB-list mailing list
omniORB-list at omniorb-support.com
http://www.omniorb-support.com/mailman/listinfo/omniorb-list
--------------------------------------------------------

If you are not an intended recipient of this e-mail, please notify the sender, delete it and do not read, act upon, print, disclose, copy, retain or redistribute it. Click here for important additional terms relating to this e-mail.     http://www.ml.com/email_terms/
--------------------------------------------------------