On Monday, June 20, 2016 at 12:19:58 PM UTC-7, Niphlod wrote: > > an hour long ? never had the necessity.... if instead we're talking about > 5-10 minutes, not at all. >
To be clear, the transfer takes about 1 minute; the timeout takes over an hour. /dps > On Monday, June 20, 2016 at 8:13:53 PM UTC+2, Dave S wrote: >> >> >> >> On Monday, June 20, 2016 at 2:24:49 AM UTC-7, Niphlod wrote: >>> >>> IMHO the problem is winsock. >>> >> >> I wouldn't be surprised if it was, but am I the only one around here with >> a winsock client? Do IE users have trouble with POSTs? >> >> /dps >> >> >>> >>> On Friday, June 17, 2016 at 11:25:30 PM UTC+2, Dave S wrote: >>>> >>>> On Friday, June 17, 2016 at 7:34:18 AM UTC-7, Niphlod wrote: >>>>> >>>>> how can it be a rocket issue when the variable is the client and not >>>>> the server ? >>>>> >>>> >>>> By not handling a common error condition? By setting the timeout too >>>> long? Not requiring a keep-alive (aka heartbeat) on an open-but-quiet >>>> connection? >>>> >>>> I haven't tried running a trace from either end, so I don't know what >>>> is happening with the handshakes other than that the connection gets >>>> established, and a long time after the client is done, we get the above. >>>> >>>> /dps >>>> >>>> >>>> >>>>> >>>>> On Friday, June 17, 2016 at 1:36:29 AM UTC+2, Dave S wrote: >>>>>> >>>>>> I'm a bad boy and still using the Rocket server that comes with >>>>>> web2py. Today I ran into an interesting case. My client code sends a >>>>>> dump >>>>>> file to the server using a POST command. On linux systems, this >>>>>> completes >>>>>> in a reasonable amount of time, gets the result (json string), and both >>>>>> sides are happy. On several Windows systems, the client reports all >>>>>> done, >>>>>> and is happy, but the server logs don't have anything and the file >>>>>> doesn't >>>>>> show up. [tick tick tick ... time passes] Oh, yes there is something! >>>>>> Everything now looks good, except that web2py.log has this from Rocket: >>>>>> >>>>>> 2016-06-16 22:10:34,846 - Rocket.Errors.Thread-5 - ERROR - Unhandled >>>>>> Error when serving connection: >>>>>> Traceback (most recent call last): >>>>>> >>>>>> File "/home/ec2-user/web2py/web2py-2.14.6/gluon/rocket.py", line >>>>>> 1337, in run >>>>>> self.run_app(conn) >>>>>> File "/home/ec2-user/web2py/web2py-2.14.6/gluon/rocket.py", line >>>>>> 1851, in run_app >>>>>> self.write(data, sections) >>>>>> File "/home/ec2-user/web2py/web2py-2.14.6/gluon/rocket.py", line >>>>>> 1766, in write >>>>>> self.send_headers(data, sections) >>>>>> File "/home/ec2-user/web2py/web2py-2.14.6/gluon/rocket.py", line >>>>>> 1750, in send_headers >>>>>> self.conn.sendall(b(header_data)) >>>>>> File "/usr/lib64/python2.7/socket.py", line 228, in meth >>>>>> return getattr(self._sock,name)(*args) >>>>>> error: [Errno 32] Broken pipe >>>>>> >>>>>> >>>>>> Timeout seems to be on the order of an hour. Linux client code uses >>>>>> libcurl, and Windows client code uses Winsock. >>>>>> The above clip is from web2py-2.14.6 running on an AWS server, from >>>>>> the zip file which has Rocket 1.2.6. >>>>>> >>>>>> Is this an issue with Rocket, or with Winsock, or each playing to the >>>>>> other's weakness? When I finally move to nginx, will these errors go >>>>>> away? >>>>>> >>>>>> /dps >>>>>> >>>>>> -- Resources: - http://web2py.com - http://web2py.com/book (Documentation) - http://github.com/web2py/web2py (Source code) - https://code.google.com/p/web2py/issues/list (Report Issues) --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
