[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
NFS hangs (again)
We're stil seeing some NFS hangs with NetBSD 5. We've applied the
PR# 39420 - stopped processes can hold locks
which may be helping, but there are still hangs.
When this happens, we see this:
nfs send error 35 for myserver:/users
Processes in "tstile" start to pile up and a handful (4-8) of processes
are in "nfssnd". Sometimes the hang goes away by iself in a few minutes.
Other times I kill the processes in "nfssnd". That appears to make the
hang go away, but it might be a coincidence.
These are TCP mounts. With UDP, we don't get the "nfs send error", but
the problem was generally worse.
We see this on both Xen domUs and directly on hardware. When it happens
on Xen, I can see that a large portion of the incoming packets are being
dropped. This doesn't seem to affect non-NFS traffic though.
As I said, this also happens on hardware, so I don't think this is a Xen
Main Index |
Thread Index |