CERN-BARC Video Conference

Wednesday 1 Nov 2006, 10:30 → 11:00 Europe/Zurich

CERN

Regular CERN-BARC Video conference on ELFms developments BARC phone: +91 22 25595228

Murthy: Hello German, My status report for video conference is as below, 1. There was a bug in TCP transport in handling large value sample. It is removed, tested and changes comitted to cvs. 2. Lemon samples encryption is successful between agent & server over TCP channel. Testing is done and server side changes are committed to CVS. The client changes are big and planning to commit them slowly in this week. UDP transport requires a complete change in buffer handling and make it similar to tcp channel. This work will be taken up now. regards murthy ------------------------ Salim: Hi German, I'm working with Miro and Dennis to finalize the format in which XML parsed data will be returned to client. Disscussed various schemes like comma separated strings, linked lists, iterators. Still putting thougts to optimize it. Also working on lemon data class methods(using xmlTextReader API). Regards, Salim ------------------------ Dinesh: Dear German & Bill , CCTracker status: The bugs reported by Vlado in CCTracker release 1.5 have been fixed and tested. These are all related to GUI of CCTracker. I will discuss the suggestion to change the publisher from Dinesh to CERN IT-FIO with Bill. This information is entered during signing the jars. The bugs reported by Bill related to updating the objects are also fixed and tested. I will now prepare the new release containing all these fixes. Best regards, Dinesh ------------------------- Discussion Miro-Murthy: ehalf case)otherwise it keeps adding the next sample in the same buffer as the msa-id is same. I did not want to disturb this setup as it is working smoothly, hence not made any changes. I just introduced a decission to switch to new buffer when ( either the msa-id changes or the metric to be sent is encryption enabled ). This setup is working smoothly in the just concluded tests of encryption over tcp channel. The thing we have to change in udp is the structure to hold the samples and a flag to remember encryption status of last metric. regards murthy Miroslav Siket wrote: > Hello Murthy, > > you have not mentioned that you were using two separate buffers for > the encrypted and non-encrypted samples. Do we really need that? Isn't > that selection whether to encrypt or not based on transport/metric > class/class level settings which should then translate into metric > level and transport destination and that it should be sufficient to > have lookup dictionary to figure out which metrics need to be > encrypted? Since you use that already to put data into two different > buffers anyway... > > Cheers, > Miro