Hey Guys, it's been a long long long time since I filed this. Unfortunately I do not have access to these servers anymore, nor do I have access to a similar solution to test this. Unfortunately I'm unable to assist any further with this bug.
Patrick ---------------------------------------- patrick <at> eefy <dot> net On Mon, Oct 5, 2009 at 11:32 PM, Ralph Janke <ubu...@janke.me.uk> wrote: > We'd like to figure out what's causing this bug for you, but we haven't > heard back from you in a while. Could you please provide the requested > information? Thanks! > > -- > rmem increase, similar to gentoo's. > https://bugs.launchpad.net/bugs/39493 > You received this bug notification because you are a direct subscriber > of the bug. > > Status in “procps” package in Ubuntu: Incomplete > > Bug description: > Hi, > > I'm not sure this's a bug per say, but it's from a recent experience i've > had, and it solved my problem. > > I've got an NFS server ( ubuntu - 2.6.15-20-amd64-server ) and when the > server is under high load the two clients ( ubuntu - 2.6.15-20-amd64-server ) > continually give the following in the messages log : > > Apr 7 16:54:43 mx01 kernel: [91655.085293] nfs: server xxxx not responding, > still trying > Apr 7 16:54:44 mx01 kernel: [91655.622991] nfs: server xxx OK > > when i say high load, i mean if i were to take the 100gb of 1k files, and > chown them all, or chmod them all ( chmod -R/chown -R ) then the server > becomes busy, and the clients seem to ''time'' out, or such... but then i > have a gentoo server ( that i'm busy phasing out in favour of the ubuntu > servers ) and it doesn't have the problem, although it is also running the > same versions of software, give or take a version. > > In the NFS performance manual on the http://nfs.sf.net site they mention > rmem, to be honest i've got no idea what it does, but i was comparing things > on the gentoo box to that on the ubuntu box in an attempt to see if it was > configuration, or possibly somthing in the binary's or kernel ( differences > on a code level, instead of a configuration level ) > > in my findings i found that the ubuntu server has the following sysctl flags > by default : > > net.core.rmem_default = 105472 > net.core.rmem_max = 131071 > > whereas the gentoo server had : > > net.core.rmem_default = 135168 > net.core.rmem_max = 131071 > > when i updated the ubuntu box to test, with the same sysctl flags for those > two ''tags'' the problem's i was having with NFS went away, i'm not sure > where i should file this bug, or who'm i should file it to, but would it be > possible for the nfs-common binary to update /etc/sysctl.conf and add the > following : > > net/core/rmem_default=135168 > net/core/rmem_max=131071 > > Thanks > > > -- rmem increase, similar to gentoo's. https://bugs.launchpad.net/bugs/39493 You received this bug notification because you are a member of Ubuntu Bugs, which is a direct subscriber. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
