[ofa-general] Ofed v1.2rc2 IPoIB

[Scott Shaw]

Hi, 
I have a small cluster setup with NFS over IPoIB device and I am seeing
a high rate of transmit timed out errors begin logged in
/var/log/messages.  What could be causing the problem and is there a
fix? 

I am using a dual port DDR Mellanox Technologies MT25208 HCA within a
DDR IB fabric.

/etc/init.d/oenibd status reports 
  HCA driver loaded
Configured devices:
ib0
Currently active devices:
ib0
The following OFED modules are loaded:
  rdma_ucm
  rdma_cm
  ib_addr
  ib_local_sa
  ib_ipoib
  ib_ipath
  ib_mthca
  ib_uverbs
  ib_umad
  ib_sa
  ib_cm
  ib_mad
  ib_core

SUSE Linux Enterprise Server 10 (x86_64)
VERSION = 10
PATCHLEVEL = 1


Jun 29 15:46:57 service2 kernel: NETDEV WATCHDOG: ib0: transmit timed
out
Jun 29 15:46:57 service2 kernel: ib0: transmit timeout: latency 1576
msecs
Jun 29 15:46:57 service2 kernel: ib0: queue stopped 1, tx_head 6355,
tx_tail 6291
Jun 29 15:46:58 service2 kernel: NETDEV WATCHDOG: ib0: transmit timed
out
Jun 29 15:46:58 service2 kernel: ib0: transmit timeout: latency 2576
msecs
Jun 29 15:46:58 service2 kernel: ib0: queue stopped 1, tx_head 6355,
tx_tail 6291
Jun 29 15:46:59 service2 kernel: NETDEV WATCHDOG: ib0: transmit timed
out
Jun 29 15:46:59 service2 kernel: ib0: transmit timeout: latency 3576
msecs
Jun 29 15:46:59 service2 kernel: ib0: queue stopped 1, tx_head 6355,
tx_tail 6291
Jun 29 15:47:00 service2 kernel: NETDEV WATCHDOG: ib0: transmit timed
out
Jun 29 15:47:00 service2 kernel: ib0: transmit timeout: latency 4576
msecs
Jun 29 15:47:00 service2 kernel: ib0: queue stopped 1, tx_head 6355,
tx_tail 6291
Jun 29 15:47:01 service2 kernel: NETDEV WATCHDOG: ib0: transmit timed
out
Jun 29 15:47:01 service2 kernel: ib0: transmit timeout: latency 5576
msecs
Jun 29 15:47:01 service2 kernel: ib0: queue stopped 1, tx_head 6355,
tx_tail 6291
Jun 29 15:47:02 service2 kernel: NETDEV WATCHDOG: ib0: transmit timed
out
Jun 29 15:47:02 service2 kernel: ib0: transmit timeout: latency 6576
msecs
Jun 29 15:47:02 service2 kernel: ib0: queue stopped 1, tx_head 6355,
tx_tail 6291
Jun 29 15:47:03 service2 kernel: NETDEV WATCHDOG: ib0: transmit timed
out
Jun 29 15:47:03 service2 kernel: ib0: transmit timeout: latency 7576
msecs
Jun 29 15:47:03 service2 kernel: ib0: queue stopped 1, tx_head 6355,
tx_tail 6291

TIA!

Scott Shaw
SILICON GRAPHICS  |  The Source of Innovation  and  Discovery
Office Ph: 734.437.6397   Cell Ph: 734.564.3832
Email:sshaw at sgi.com     http://www.sgi.com