[rabbitmq-discuss] RabbitMQ running at 100% CPU.
chime at mu.dk
Tue Mar 25 14:55:10 GMT 2008
Matthias (and others)
Could this http://www.nabble.com/Unix-spawn-driver-does-not-close-all-file-descriptors-in-child-process-td15796733.html
explain some of the weirdness we have seen on OS X?
On Mar 13, 2008, at 16:36 , Michael Arnoldus wrote:
> Seen on several Mac's here, but the one where we have the problem
> right now is a Quad-core 2,8 GHz Mac Pro. Also seen on a dual core.
> 2 GB Mem. Running Mac OS X 10.5.2 (Leopard). Erlang 11B-5, Rabbit
> I actually have a system running with the problem, so if I can help
> in any way tracing the problem down, let me know.
> Yes, Mac OS X has dtrace.
> It seems to me like it should be possible to find an possible
> infinite loop on writev error writing to a socket, either in the
> erlang code or the rabbit code (but then again - what do I know :-)
> Let me know if you have any suggestions on where I should start
> On Mar 13, 2008, at 16:11 , Matthias Radestock wrote:
>> Michael Arnoldus wrote:
>>> Thank you for your suggestion. There is no strace on Mac OS X, but
>>> I did find a way to see what the C-program was doing (see below).
>> I gather there's ktrace and/or dtrace.
>>> Most of the threads are just waiting (as expected) but a single
>>> one seems always to be in some kind of tcp_send_error.
>> That is interesting. Tony has seen similar symptoms on his MacBook
>> here, but so far we have failed to reproduce the problem anywhere
>> Please let us know as much about your system as possible. Exact OS
>> version, Erlang/OTP version, RabbitMQ version, number and exact
>> type of processors, memory configuration, and anything else you can
>> think of that might be relevant.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 1912 bytes
Desc: not available
Url : http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20080325/99e44bd4/attachment.bin
More information about the rabbitmq-discuss