Skip to main content

Enable Fast I/O Failure for FC adapt

AIX 5.2 supports the Fast I/O Failure feature for Fibre Channel devices in case of link events in a switched environment. If the FC adapter driver detects a link event (for example, a lost link between a storage device and a switch), the FC adapter driver waits a short period of time, about 15 seconds, to allow the SAN fabric to stabilize. At this point, if the FC adapter driver detects that the device is not on the SAN fabric, it begins failing all I/O requests at the adapter driver level. Any new I/O request, or future retries of the previously failed I/Os, are failed immediately by the adapter, until the adapter driver detects that the device has rejoined the SAN fabric.


Fast I/O Failure is controlled by a new fscsi device attribute, fc_err_recov. The default setting for this attribute is delayed_fail, which is the I/O failure behavior that has existed in previous versions of AIX. Setting this attribute to fast_fail enables Fast I/O Failure, as shown in Example 3-12.

Example 3-12 FC adapter fast_fail option

{node1:root}/_> chdev -l fscsi0 -a fc_err_recov=fast_fail

The fscsi0 device is used for storage data transfer and is a child device of fcs0 (the FC adapter):

+ fcs0 U0.1-P1-I4/Q1 FC Adapter* fscsi0 U0.1-P1-I4/Q1 FC SCSI I/O Controller Protocol DeviceFast fail logic is invoked when the adapter driver receives an indication from the switch that there has been a link event involving a remote storage device port via a Registered State Change Notification (RSCN) from the switch.

Fast I/O Failure may be desirable in situations where multipath software is being used. Setting fc_err_recov to fast_fail may decrease the I/O fail times due to link loss between the storage device and switch and allow faster failover to alternate paths.

In single-path configurations, especially in configurations with a single path to a paging device, the default delayed_fail setting is recommended.

Fast I/O Failure is only supported in a switched environment. It is not supported in arbitrated loop environments, including public loop.

We configured Fast I/O Failure since we were using the Subsystem Device Driver, two adapters per node and two paths to the storage subsystem.

The requirements for Fast I/O Failure support are the following:

FC 6227 adapter firmware - level 3.22A1 or greater

FC 6228 adapter firmware - level 3.82A1 or greater

FC 6239 adapter firmware - all firmware levels

Comments

Post a Comment

Popular posts from this blog

Configuring the virtual path devices

We recommend that you start with a "fresh" disk configuration, so it is a good idea to delete all previously configured FC adapters and their child (disk) devices. On node1, we checked which disks are still defined: {node1:root}/-> lspvhdisk0 0022be2ab1cd11ac rootvg activehdisk1 0022be2a3d02ead0 Nonehdisk2 0022be2a4cbbafd8 Nonehdisk3 none NoneThese are the internal SCSI disk drives {node1:root}/-> lscfg grep disk+ hdisk3 U1.9-P2/Z2-A8 16 Bit LVD SCSI Disk Drive (36400 MB)+ hdisk2 U1.9-P2/Z1-A8 16 Bit LVD SCSI Disk Drive (36400 MB)+ hdisk1 U1.9-P1/Z2-A8 16 Bit LVD SCSI Disk Drive (36400 MB)+ hdisk0 U1.9-P1/Z1-A8 16 Bit LVD SCSI Disk Drive (36400 MB)In order to include the ESS disks, run the configuration manager on each node: {node1:root}/-> cfgmgr -vSince ESS was configured with two host paths for each node (node1a and node1b), this results in two hdisks on the nodes. Actually, those two logical hdisks represent the same physical disk, accessed via the two ...
Post a Comment
Read more

The HACMP clstat command does not work

Clinfo daemon started? The clstat command retrieves the information from the clinfo daemon. This daemon is not started automatically. It is an option in the SMIT HACMP startup menu (smitty clstart). If it is not running, stop HACMP on this node, and start it up again with clinfo activated. To check the status of the HACMP processes: lssrc -a | grep -E "ES | svcs" SNMP version in AIX 5L version 5.2 AIX 5L version 5.2 uses SNMP version 3 agents, whereas HACMP uses a SNMP version 1 configuration. Both clinfo and CSPOC require SNMP version 1. With the default SNMP agent, the clinfo daemon is not able to supply cluster information, thus the clinfo command will not work properly. Run the script provided in Example 3-38 to switch to version 1 agents. Example 3-38 Switch SNMP agents to version 1 #!/bin/kshLOG=/var/adm/chg_snmpd.logtouch $LOG/usr/es/sbin/cluster/utilities/clstop -N -g >> $LOG 2>&1sleep 5for i in muxatmd aixmibd snmpmibd hostmibd snmpddosto...
Post a Comment
Read more

Typical Oracle RAC configurations

All configurations are based on the following building blocks: Hardware Server nodes Storage Networking Software Operating system Cluster software Oracle RAC (application) Application architecture Oracle9i RAC on RAW devices is based on a shared disk architecture. Figure 2-1 shows a two-node cluster. The lower solid line is the primary Oracle interconnect, the middle dashed line is the secondary Oracle interconnect. For high availability, both these networks should be defined in the HACMP as "private". HACMP/ESCRM provides Oracle9i RAC with the infrastructure for concurrent access to disks. Although HACMP provides concurrent access and a disk locking mechanism, this mechanism is not used. Oracle, instead, provides its own locking mechanism for concurrent data access, integrity, and consistency. Volume groups are varied on all the nodes, thus ensuring short failover time. This type of concurrent access can only be provided for RAW ...
Post a Comment
Read more