[Xorp-users] xorpsh hanging on 'show pim join'.

[Pavlin Radoslavov]

Ian Batterbee <ian.batterbee at aut.ac.nz> wrote:

> Hi, I'm seeing some unexpected behaviour through xorpsh, in that the 
> backend module appears to become unresponsive.
> 
> If I issue "show pim join" repeatedly, the first few requests succeed, 
> then subsequent commands never return - after about 15 (guess) seconds, 
> the xorp_rtrmgr process says:
> 
> "[ 2007/07/24 17:52:10  ERROR xorp_rtrmgr:20556 FINDER +85 
> finder_xrl_queue.hh dispatch_cb ] Sent xrl got response 211 Reply timed out"
> 
> but xorpsh continues to wait until I hit ^C, at which point the rtrmgr says:
> 
> [ 2007/07/24 17:52:56  ERROR xorp_pimsm4:20566 LIBXORP +213 
> buffered_asyncio.cc io_event ] read error 104
> [ 2007/07/24 17:52:56  ERROR xorp_pimsm4:20566 XRL +159 xrl_pf_stcp.cc 
> read_event ] Read failed (error = 104)
> [ 2007/07/24 17:52:56  ERROR xorp_pimsm4:20566 XRL +338 xrl_pf_stcp.cc 
> die ] STCPRequestHandler died: read error
> 
> ...presumably in response to me aborting the request.
> 
> At that point, I can *occasionally* press [up] in xorpsh and the command 
> works, but most of the time it hangs.
> 
> if I shut xorp_rtrmgr down, and then restart it, things are okay again 
> for a few minutes, and then the problem begins once more.
> 
> xorp is speaking PIMv4 to a cisco router which connects to a 
> education/research network, which has a lot of multicast traffic on it. 
> Could the amount of sources could potentially be the problem ?
> 
> If I simply start xorp_rtrmgr up, and then wait 5 minutes, the first 
> time I issue 'show pim join', the output hangs.
> 
> It *may* also have something to do with me piping the output through "| 
> no-page " or "| find 233.0.76.26".. I'm not sure if that's related but I 
> thought I better mention it in case it is.

Do you always use the "| no-page" or "| find 233.0.76.26" piping or
it hangs even on "show pim join" only?

Could you try "show pim join | count" to get some idea about the
size of the output.
If this fails, can you provide some estimation about the number of
(S,G) and (*,G) PIM-SM entries inside the XORP router.
Finally, could you confirm that running other commands like
"show pim interface" don't create such problem.

FYI, the current mechanism for obtaining the "show ..." output from
the multicast-related processes is very hackish and needs to be
replaced.
Hopefully, after the replacement the problem you are seeing will go
away, but it is good to know what is the source of the problem.
My suspicion is the output size is relatively large, which the
current mechanism probably doesn't handle well.

Thanks,
Pavlin

> _______________________________________________
> Xorp-users mailing list
> Xorp-users at xorp.org
> http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-users