User guide

DataDirect Networks SFA™ OS 1.5.3.0 Release Notes Revision B | 12
5.2 SFA12K-20E
On SFA12K-20E systems, if you choose to use the enhanced VM network management
(SR-IOV), there is a problem which prevents binding the app IOC associated with the
second physical interface on the SFA controller. If you bind the second interface, the VM
may fail to start properly.
To avoid this problem, use the CLI to bind these IOCs to stacks (VMs):
app bind stack 0 ioc 01792
app bind stack 32768 ioc 34560
NOTE: The SFA controller can have its management network configured on either eth0
or eth1. However, the VMs can only make use of SR-IOV enhanced VM
management networking from controller eth0. Do not use the IOCs associated
with the controller’s eth1 (01793 or 34561).
5.3 SFA12K and SFA10K
• Issuing a DELETE CONFIG command to a SFA12K-20E with IB removes the
discovered initiators but leaves the channels up. Host computers and the SFA
may indicate valid connections; however, they will not function until the SFA is
restarted.
• QLogic IB HCA and IB switch products may fail to negotiate QDR speeds with an
IB-attached SFA product. This is not an issue in SFA OS, but is a known issue in
the IB driver.
• The SS7000 enclosure does not support the SET SLOT xx POWER OFF
command with enclosure firmware versions prior to 05.02.01.
• In the SS7000, when removing an I/O module or cable under load, it is possible
that one or more disks could be marked failed which would change the
redundancy of the storage pools. This is due to in-flight I/O being timed out and
retried before the IOC reports the drive as missing. There is a high probability
that the newly failed drives are fine and can be returned to service by clearing the
failed state and reassigning them to pools.
Use the following procedure to clear a failed drive and reassign it back to a pool:
show unassigned failed

Shows which PD have failed; get drive’s
<id>
show pool
Shows which pools are involved
clear pd <id> failed

<id> is from show unassigned output
assign pd <id> to pool-id set
Pool-ID is shown in the show pool output
NOTE: If the pool sparing policy is set to 'SWAP', the SFA will typically auto
rebuild the drives, using fractional or full rebuild depending on the
duration of the interruption.
• In an InfiniBand switch environment, there is a small chance that if a cable
between the switch and the controller is pulled, both the physical link and the
logical link will be lost. This has occurred with a Mellanox 6025F Switch and a
Mellanox HCA.