On Tue, Nov 20, 2007 at 03:53:34PM +0100, Mark Hunting wrote: > Dave Dykstra wrote: > > Not to discourage you, but once you get past that you may also find you > > have problems with getting locks to survive server fail-overs. See > > my last posting I had on the subject was over 2 years ago: > > > > http://lists.community.tummy.com/pipermail/linux-ha/2005-July/015357.html > > > > Nobody replied so I don't think very many people have tried doing locks > > on linuxHA NFS. I still don't know what to do about the discussion > > on the HaNFS web page, and for the last year and a half I haven't > > even worked at the company where I had a linuxHA NFS setup. > > > Dave, I wrote earlier that I do not have the problems you had. But after > reading your original mail better and doing more tests I now see that I > have exactly the same problem. This is really bad, now I have to choose > between disabling file locking completely or take the risk of locking > failures during take-over. I think I will choose the first, as it will > also 'solve' my first problem. > > Mark
Well at least I'm glad to hear that someone else could confirm my experience, although it's too bad that there's not yet a solution to make it work even after all this time. I guess I will update the wiki page to say that locking doesn't work across failovers at least under certain conditions which I will list, recommend that people mount NFS with the "nolock" option, and move the old info about locking to a backup page in case it's still of use to someone. Please let me know what kernel you tested with, and whether you were using NFS-over-UDP or NFS-over-TCP. - Dave _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
