The working one has lots of interleaved connections. Can you make sure that everything using katcp is off, then start capturing, then start katcp again.
--Andy On Tue, May 21, 2013 at 2:41 PM, Haoxuan Zheng <jef...@mit.edu> wrote: > Hi Andy, > Here are the wireshark files from the working computer and non-working > computer. The working one does have many more [PSH, ACK] communications. > > Thanks a lot for your help! > Jeff > ________________________________________ > From: aml...@gmail.com [aml...@gmail.com] on behalf of Andrew Lutomirski > [l...@mit.edu] > Sent: Tuesday, May 21, 2013 5:12 PM > To: Haoxuan Zheng > Cc: David MacMahon; casper@lists.berkeley.edu > Subject: Re: [casper] Python katcp cannot connect FPGA > > On Tue, May 21, 2013 at 1:36 PM, Haoxuan Zheng <jef...@mit.edu> wrote: >> Hi David, >> I tried everything you suggested, and here's what I got: >> 1) I don't see any new messages with dmesg on ROACH2 after the attempted >> connection; maybe I did something wrong? I have never used dmesg before. >> 2) I attached the wireshark screen shot comparing the working computer and >> the non-working on. Basically the non-working computer goes through all the >> processes that the working computer does, and then actively closes the >> connection with [FIN, ACK]. Does this give any clues? >> 3) The non-working computer had 0.7.3 which was newer than the working >> 0.7.1, but uninstalling and installing 0.7.1 did not solve the problem. >> > > It might be worth doing a "Follow TCP conversation" or sending a .pcap > file from Wireshark. > > --Andy > >> >> Thanks a lot! >> Jeff >> >> ________________________________________ >> From: David MacMahon [dav...@berkeley.edu] on behalf of David MacMahon >> [dav...@astro.berkeley.edu] >> Sent: Monday, May 20, 2013 12:58 PM >> To: Haoxuan Zheng >> Cc: casper@lists.berkeley.edu >> Subject: Re: [casper] Python katcp cannot connect FPGA >> >> Hi, Jeff, >> >> That sounds like a weird one! Here are some ideas to try: >> >> 1) Check the output of "dmesg" on the roach2 after the problem occurs. >> >> 2) Watch the network traffic between the problematic machine and the roach2 >> using tcpdump or wireshark. >> >> 3) Verify that the problematic computer is using the same python katcp >> version as the working computers. >> >> Please send a message to the list once you figure it out. You'll likely >> make some CASPER mailing list archive reader very happy in the future! :-) >> >> Hope this helps, >> Dave >> >> On May 20, 2013, at 9:01 AM, Haoxuan Zheng wrote: >> >>> Hi CASPER, >>> Sorry for the spam, but since my original message was sent at the worst >>> hour of the week (Friday evening), I would like to give it another try: >>> >>> Our lab just got a new server which will be our data taking computer, and >>> it has trouble connecting to our ROACH 2 through python katcp. Whenever I >>> do fpga = corr.katcp_wrapper.FpgaClient('10.0.0.200', 7147, timeout = 60), >>> fpga.is_connected() always say False. Here are some clues: >>> >>> 1) When I run the above command, I see the following on a separate telnet >>> 7147 connection on the ROACH2 (which runs fine), and the >>> client-disconnected message comes right after the client-connected message: >>> >>> #log info 946685931331 raw new\_client\_connection\_10.0.0.11:44001 >>> #client-connected 10.0.0.11:44001 >>> #client-disconnected 10.0.0.11:44001 >>> >>> 2) The new computer connects to the ROACH1s fine with python katcp. >>> >>> 3) All other computers connect to the ROACH2s fine with python katcp. >>> >>> I'm not sure if anyone else has experienced similar issues? Any help would >>> be greatly appreciated! >>> >>> Thanks a lot! >>> Jeff >>