|
I guess that "over NFS" makes the difference here. Spooling over NFS is not the best thing to do performance-wise. You might want or need to do it because you have a fileserver which is highly available, of course. Cheers, Fritz On Fri, 8 Apr 2011 at 10:32am, Stephen Dennis wroteHi MarkI am with Chris. Classic meets large scale performance and reliability requirements. I do not know any compelling reason to use BerkleyDB.Interesting. I did some benchmarks a year ago on our production cluster (see <http://markmail.org/message/xni5kqj2h3t6qjkd>) and saw a definite difference between classic and BDB when both were run over NFS. Based on those (and the fact that misbehaving users submitting several thousands of jobs have crashed our scheduler not infrequently), I've been leaning towards moving our install to BDB. Are there other tunables I should be looking at? --
|
Fritz Ferstl | CTO and Business
Development, EMEA
_______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
