Den 2012-03-23 14:18:06 skrev Lluís Batlle i Rossell <[email protected]>:
> On Thu, Mar 22, 2012 at 03:20:46PM +0100, Eelco Dolstra wrote: >> On 22/03/12 14:02, Lluís Batlle i Rossell wrote: >> >> > It is still broken, right? >> >> The test passed for me, but there still seems to be some random race >> condition (?): >> >> http://hydra.nixos.org/build/2315700 >> >> In particular: >> >> > server# [ 133.310848] lockd: rejected NSM callback from 127.0.0.1, >> port=40233 >> > server# Mar 22 14:13:28 server rpc.statd[1837]: nsm_parse_reply: >> [0x4f6f9fd2] RPC accept status 5 >> >> This seems to suggest that statd is sometimes not using a privileged >> port to >> talk to the kernel. Maybe passing --outgoing-port to statd will help. > > I don't know at all. The analysis does not look like it should have > symptoms similar to > a race. > > Rickard, our NFS expert, any clue? I think Eelco is right about the privileged port, I wonder why statd uses port 40233? Does this only happen in the automated test? / Rickard _______________________________________________ nix-dev mailing list [email protected] http://lists.science.uu.nl/mailman/listinfo/nix-dev
