> From: [email protected] [mailto:[email protected]] > On Behalf Of Doug Hughes > > Ah, I think I know exactly what this (above) is. Linux (CentOS) does > not fully implement the READDIRPLUS NFS RPC call, but Solaris does. When > the client asks for READDIRPLUS, the server fills in the packet with a > large number of replies and sends them in bulk. When the client asks for > READDIR, every single entry is enumerated as a single RPC which means > 7000 RTTs. If you are using ls -l or ls -color, automatically double > that to 14000 RTTs because it will be doing getxattr and/or stat calls > for every entry. ouch!
So... Can you just upgrade the version of the NFS client, or something like that? Perhaps download an auxiliary package from EPEL or RPMForge, or something which isn't part of the mainstream RHEL/Centos repositories? I certainly know I can't settle for the default autofs4 that's included with the distro. Always upgrade to autofs5. And install python2.6. All of which is unrelated, except to say a certain amount of upgrading above and beyond what's included in the distro is absolutely necessary if you want a system that doesn't suck... Based on my quick search, it looks like .... kernel, glibc, portmap, nfs4-acl-tools, nfsutils... Are the suspects. I don't see any significantly newer nfsutils or nfs4-acl-tools for centos5. That's where I stopped. Didn't look any further for portmap, glibc, or what implications it may have if you upgrade them. _______________________________________________ Tech mailing list [email protected] https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
