Hi Elvin,
Then this is a bug in the client. I had a long discussion about this with
Lukasz. The final implementation was supposed not close the TCP
connection if there was an outstanding response. This, of course, does
not apply when the server sent a simple "wait" as there is no response
outstanding and the request can be resent after the wait interval. This
is not the case with "waitresp" as that implies that the response will be
sent later. The endsess failure indicates that there was atleast one
request still active when the connection was closed. So, this is indicatve
of an error. Of course, loosing the opaque information on the retry is
another problem.
BTW does Castor send waitresp responses (i.e. the ofs plugin returns
SFS_STARTED)?
Andy
On Thu, 17 Sep 2015, Elvin Sindrilaru wrote:
>
> Yes, that's exactly what I also see in the logs.
>
> 150916 17:41:32 time=1442418092.137614 func=open level=INFO logid=677fda4e-5c89-11e5-8736-001b21551dbc [log in to unmask]:1094 tid=140532264986368 source=XrdxCastor2Ofs:448 tident=alicedaq.8138:353@aldaqtdsm03 path=/castor/cern.ch/alice/raw/technical/2015/09/16/16/15000236105030.1319.root, opaque=castor.accessop=3&castor.client_sec_gid=1395&castor.client_sec_uid=13798&castor.exptime=1442418152&castor.id=alicedaq.8138:505@aldaqtdsm03&castor.manager=c2alicesrv201.cern.ch:1094&[log in to unmask]&castor.pfn2=0:15511:1ec25b08-f4d2-511b-e053-9d08100ab58e&castor.sfn=/castor/cern.ch/alice/raw/technical/2015/09/16/16/15000236105030.1319.root&castor.signature=We5WNRa9z64XHnm2oiucnUo0FCcjxh0r5F9oLhaXq1CJ/6FkfyqYJ7I6UGAlzX0Vx7jS+3CrEOxDtkMJsmrmhQ==&castor.txtype=user&mode=644&oss.asize=1865279545, isRW=1, open_mode=4102, file_ptr=0x7fd0d27a4f10
> 150916 17:43:22 time=1442418202.152684 func=open level=INFO logid=a912d31c-5c89-11e5-bf06-001b21551dbc [log in to unmask]:1094 tid=140534474995456 source=XrdxCastor2Ofs:448 tident=alicedaq.8138:441@aldaqtdsm03 path=/castor/cern.ch/alice/raw/technical/2015/09/16/16/15000236105030.1319.root, opaque=mode=644&oss.asize=1865279545, isRW=1, open_mode=4002, file_ptr=0x7fd0d24ed010
>
> So, it appears that the DefaultConnectionWindow = 120 was the one responsible for the retry.
> I'm digging deeper ...
>
>
> ---
> Reply to this email directly or view it on GitHub:
> https://github.com/xrootd/xrootd/issues/287#issuecomment-141016534
—
Reply to this email directly or view it on GitHub.
Use REPLY-ALL to reply to list
To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1