zfs, nfs and zil

[Jeremy Chadwick]

On Tue, Mar 29, 2011 at 09:37:34PM +0200, Claus Guttesen wrote:
> >> All ip-activitiy, dns, ssh etc. on the interface (ip-address) that is
> >> mounted. So whenever I test I log in to the server using the other
> >> ip-address and whenever it stops responding to ping or my screen
> >> session stops I reload the interfaces and the traffic resumes.
> >
> > Are the two NICs the same brand/driver/board? If not, could you try to
> > switch the networks on them and test if the problem persists? It would also
> > be nice if you could tell which driver/chip the NIC with the problem is.
> 
> It's a supermicro server with dual igb-nics on the mainboard. I did
> try other cables and other ports on the switch (hp procurve). I don't
> have access to the server atm. but I'll get the chip-info in a day or
> two. Both ip-addresses are on same subnet.
> 
> Speed is autonegotiated to 1 GB full duplex on server and switch. No
> errors detected using netstat or on switch.

Something tells me (gut feeling) the virtualisation part is probably
somehow responsible for the problem, quote:

> I'm mounting the FreeBSD-server from a couple of vmware esxi 4.1
> servers using nfs, but when there is alot of i/o the server becomes
> unresponsive, easily triggered by installing ie. ms-sql. The server
> itself is up but is not reachable from the network. When I take the
> nic down and up again connection to the network is reestablished
> (ip-wise).

Can you remove ESXi from the picture and see if the problem continues?
If it does, we should probably pull Jack Vogel into the discussion, as
Supermicro predominantly uses Intel-based NICs.  He'll need output from
"pciconf -lvcb" and "dmesg" to assist.

But before pulling him in, please see if you can remove ESXi from the
picture.

-- 
| Jeremy Chadwick                                   jdc at parodius.com |
| Parodius Networking                       http://www.parodius.com/ |
| UNIX Systems Administrator                  Mountain View, CA, USA |
| Making life hard for others since 1977.               PGP 4BD6C0CB |