RHEL6 is autotuning. the boxes i am running have largish memory (128GB) and have plenty free.
i looked through the tuning page, but it basically says if autotuning is on, then there is not much to do. i added some debugging, and it seems like we're asking for a lot. one process asked for and received 256K. the process that crashes has got 256K on one file descriptor, and is failing while trying to push from 256K to 512K on another file descriptor. does this seem reasonable? i can't tell from teh code; it seems like the code is only sending 48 bytes. this smells like the thing receiving it is not working fast enough; no? this is happenning while i am starting up teh processes; is there some setup stuff going on that is relatively slow? andrew On Jun 14, 2011, at 10:15 AM, Pieter Hintjens wrote: > On Tue, Jun 14, 2011 at 3:01 PM, Andrew Hume <[email protected]> wrote: > >> Assertion failed: new_sndbuf > old_sndbuf (mailbox.cpp:183) > > The assertion is annoyingly useless, IMO the code should at least say > something like "Insufficient socket buffer space available" before > dying. > > The problem is that the OS isn't providing enough buffer space for the > socketpair used for internal mailbox communications. Linux does set > its limits automatically but RHEL6 might be not doing this. Here's a > page that may help: > http://www.psc.edu/networking/projects/tcptune/#Linux > > The problem is more typically seen on BSD systems (Mac OS/X and > FreeBSD): http://www.zeromq.org/docs:tuning-zeromq > > -Pieter > _______________________________________________ > zeromq-dev mailing list > [email protected] > http://lists.zeromq.org/mailman/listinfo/zeromq-dev ------------------ Andrew Hume (best -> Telework) +1 623-551-2845 [email protected] (Work) +1 973-236-2014 AT&T Labs - Research; member of USENIX and LOPSA
_______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
