Print

Print


Hi Andreas,

I don't see any immediate problem with the config file (though you are 
using the old format and it could be simplified). However, the problem is 
that the data server olbd is waiting for the xrootd to come up and it 
appears to not be doing that. What missing below is the corresponding 
xrootd log file for each server. That would likely explain the problem.

Andy

On Fri, 2 Nov 2007, Andreas Petzold wrote:

> 	Hi,
>
> I want to setup a redirector and dataserver using the following config file:
>
> all.manager lobsang.neubert23.net 3121
> all.role manager if lobsang.neubert23.net
> all.role server if rincewind.neubert23.net
> olb.allow host rincewind.neubert23.net
> olb.port 3121
> olb.path w /home/xrootd/testdata
> xrd.protocol xrootd *
> xrootd.export /home/xrootd/testdata
> xrootd.async off
> xrd.adminpath /home/xrootd/xrdadmin
> olb.adminpath /home/xrootd/olbadmin
> olb.trace all
>
> Are there any obvious problems with this configuration?
>
> The problem is, that while the manager olbd and the server olbd start, they
> never communicate with each other. Here are the log files:
>
> Manager log:
>
> 071102 21:03:25 001 (c) 2004 Stanford University/SLAC xrd version
> 20070321-1251p1
> 071102 21:03:25 001 olbd [log in to unmask] initialization started.
> 071102 21:03:25 001 Using configuration file opt/xrootd/etc/xrootd.cfg
> 071102 21:03:25 001 Maximum number of connections is 1024
> 071102 21:03:25 001 (c) 2006 Stanford University/SLAC olbd.
> 071102 21:03:25 001 [log in to unmask] phase 1 initialization started.
> 071102 21:03:25 001 [log in to unmask] phase 1 manager initialization
> suceeded.
> 071102 21:03:25 001 [log in to unmask] phase 2 manager initialization
> started.
> 071102 21:03:25 001 olb_Config: Round robin scheduling in effect.
> 071102 21:03:25 001 [log in to unmask] phase 2 manager initialization
> completed.
> 071102 21:03:25 001 olbd [log in to unmask]:3121 initialization
> completed.
> 071102 21:04:55 4570 olb_Config: manager service enabled.
>
> Server log:
>
> 071102 21:03:46 001 (c) 2004 Stanford University/SLAC xrd version
> 20070321-1251p1
> 071102 21:03:46 001 olbd [log in to unmask] initialization started.
> 071102 21:03:46 001 Using configuration file opt/xrootd/etc/xrootd.cfg
> 071102 21:03:46 001 Maximum number of connections is 1024
> 071102 21:03:46 001 XrdSched: scheduling underused thread monitor in 780
> seconds
> 071102 21:03:46 001 XrdSched: Starting with 2 workers
> 071102 21:03:46 001 XrdLink: Allocating 16 link objects at a time
> 071102 21:03:46 001 XrdPoll: Starting poller 0
> 071102 21:03:46 001 XrdPoll: Starting poller 1
> 071102 21:03:46 001 XrdPoll: Starting poller 2
> 071102 21:03:46 001 XrdProtocol: getting port from protocol olbd
> 071102 21:03:46 001 (c) 2006 Stanford University/SLAC olbd.
> 071102 21:03:46 001 [log in to unmask] phase 1 initialization
> started.
> 071102 21:03:46 001 [log in to unmask] phase 1 server initialization
> suceeded.
> 071102 21:03:46 001 XrdProtocol: getting protocol object olbd
> 071102 21:03:46 001 [log in to unmask] phase 2 server initialization
> started.
> 071102 21:03:46 001 calcSpace New fs info; maxfree=23856404K totfree=23856404K
> 071102 21:03:46 001 olb_Meter: Found 1 filesystem(s); 22G total bytes free;
> 22G available
> 071102 21:03:46 001 [log in to unmask] phase 2 server initialization
> completed.
> 071102 21:03:46 10383 XrdSched: running olbd startup inq=0
> 071102 21:03:46 10383 olb_Start: Waiting for primary server to login.
> 071102 21:03:46 001 olbd [log in to unmask]:54267 initialization
> completed.
> 071102 21:03:46 001 XrdSched: scheduling midnight runner in 10574 seconds
> 071102 21:04:22 10383 calcSpace New fs info; maxfree=23856400K
> totfree=23856400K
> 071102 21:04:46 10383 calcSpace New fs info; maxfree=23856324K
> totfree=23856324K
> 071102 21:05:46 10383 calcSpace New fs info; maxfree=23856320K
> totfree=23856320K
>
> There is no filewall between the two hosts and when I us nc to connect to port
> 3121 on the manager I see a log message about the connection attempt in the
> master log file.
>
> I don't have an idea what I could be doing wrong. Any advice?
>
> 	Cheers,
>
> 		Andreas
>