 |
 |
Hi Marcus, The only other time I've seen this happen (I haven't yet encountered a faulty IPv6 stack like Matevz has) is due to asymmetric routing. The connect works fine but the data packets might be routed on a different path (i.e. not the initial connection path), in which case it looks like nothing has been sent. That said, what version of the MM and Manager are you running? Also, are there any unusual complaints in the Manager's log? Andy On Fri, 9 Jan 2015, Marcus Ebert wrote: > Thanks Matevz! > > Unfortunately, it doesn't solve the problem. I added that option to > xrootd and cmsd on the manger and metamanager site, but it's still the > same. > > > Cheers, > Marcus > > On Fri, Jan 9, 2015 at 9:34 AM, Matevz Tadel <[log in to unmask]> wrote: >> Hi, >> >> I had a similar problem recently ... turned out it was lousy ipv6 setup >> somewhere between the manager and meta manager. By adding '-I v4' to >> cmsd/xrootd options on the manager it started to work ok. >> >> Matevz >> >> >> On 1/9/15 9:16 AM, Marcus Ebert wrote: >>> >>> Hi all, >>> >>> I set up a meta manager to which different managers subscribe. This >>> works fine for all I set up so far at different sites, except for one >>> site. >>> When I start cmsd there I see in the log of the local manager which >>> tries to connect to the meta manager METAMANAGER (hostname remvoed >>> here, but shown correctly in the log): >>> 50109 11:48:46 22893 XrdInet: Connected to METAMANAGER:3121 >>> 150109 11:48:56 22893 manager.0:23@METAMANAGER XrdLink: No RecvAll() >>> data; errno=0 >>> 150109 11:48:56 22893 Login: METAMANAGER login failed; rejected >>> 150109 11:48:56 22893 Manager: manager.0:23@METAMANGER removed; lost >>> connection >>> >>> This repeats infinitely until cmsd is stopped and the same messages I >>> receive if I try to have the local file server directly to connect to >>> any other manager on any other site. >>> In the log of the meta manager I see messages about the manager >>> MANAGER trying to connect, like: >>> 150109 03:04:12 29463 XrdInet: Accepted connection from 23@MANAGER >>> 150109 03:04:12 29460 XrdSched: running main accept inq=0 >>> 150109 03:04:13 29463 XrdProtocol: ?:23@MANAGER terminated login not >>> received >>> >>> The same shows up if the server cmsd tries to connect to any non local >>> manager in that manager log. >>> >>> Does anyone know what could be the reason for that behavior or what >>> should be checked? The "-d" option also doesn't give any more >>> information, "trace all" is already used, and this specific behavior >>> happens only for the installation on that single machine while >>> machines on other sites can connect fine. >>> >>> >>> Thanks, >>> Marcus >>> -- >>> >>> ######################################################################## >>> Use REPLY-ALL to reply to list >>> >>> To unsubscribe from the XROOTD-L list, click the following link: >>> https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-L&A=1 >>> >> > > -- > > ######################################################################## > Use REPLY-ALL to reply to list > > To unsubscribe from the XROOTD-L list, click the following link: > https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-L&A=1 > ######################################################################## Use REPLY-ALL to reply to list To unsubscribe from the XROOTD-L list, click the following link: https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-L&A=1