Help

○ Monitor the port's counters for non-zero values for tim_txcrd_z or time transmission credits are zero. This means the FC
port wanted to transmit a FC packet, but did not have sufficient buffer credits to so. Any non-zero value in this category
implies performance issues on the WAN link.
○ If FCIP gateway devices are used between metro node clusters, ensure that the FCIP tunnel is configured properly.
● For Brocade FCIP switches:
○ Check for bandwidth rate limiting setting on the tunnel. See the portshow fciptunnel. command. Verify the values
for Minimum/Maximum Communication Rate are not causing a bottleneck.
○ Check for improper QoS settings on the tunnel. From the portshow fciptunnel command output, check the values
for QoS Percentages. Note that only if QoS has been set on the LAN facing Fibre Channel ports will QoS settings affect
the fciptunnel settings.
Corrective actions
● Check the WAN devices for improperly configured ports, link errors, packet loss, QoS limitations, or large observed round-
trip-times outside of supported metro node specs.
● Check the metro node WAN ports for issues.
● Follow the WAN products' performance best practices.
● If you suspect that you are not achieving optimal inter-cluster WAN throughput, verify the available inter-cluster bandwidth.
● Find out from your WAN provider if anything has changed recently.
● For remote read or remote write operations are limited to the remote cluster's storage-array's performance. Make sure that
the back-end performance on your remote cluster is not your bottleneck.
● From the WAN endpoint devices, check the WAN link latency statistics for irregular round-trip times.
Corrective actions for Metro-IP
● Check the network's maximum supported MTU (maximum transmission unit) size. Standard networks support 1500 byte
MTUs. Jumbo frames can support up to about 9000 byte MTUs, which typically result in faster throughputs and consumes
fewer CPU cycles. Jumbo frames however must be explicitly enabled on every device in the network path
● Check the port-group's socket buffer size (socket-buf-size in /clusters/cluster-#/cluster-connectivity/option-sets/
optionset-com-#/). The default is set to 1MB. The optimal value for this the network's delay-bandwidth product which
is the latency or delay of the network multiplied by the available bandwidth, which is the amount of data required to be
outstanding to fully utilize the network.
● If this value is too large for a low bandwidth connection, there will be connection keep alive timeouts and possibly cluster
departure events. Lower the socket-buf-size in this case.
● If this value is too low for a high bandwidth connection, there won't be enough in-flight data outstanding on the WAN
link to fully utilize it's available capacity, and the observed performance will be much less than expected. Increase the
socket-buf-size to the minimum delay-bandwidth product.
Viewing the WAN Latency chart
1. From the GUI main menu, click Performance.
2. In the Performance Dashboard, select the tab in which you want to display the WAN Latency chart (or create a custom
tab).
3. Click +Add Content.
4. Click the WAN Latency icon.
Rebuild Status dashboard
The Rebuild Status dashboard displays the status of any rebuilds or migration operations that are running on your metro node
system. Rebuilds synchronize data from a source drive to a target drive. When differences arise between legs of a RAID-1 device
(local or distributed), a rebuild updates the out-of-date leg.
NOTE:
The dashboard displays data only for the directors in the cluster to which you are currently connected. To
simultaneously view Rebuild Status for another cluster, open a second browser session and connect to the second cluster.
Monitoring the system 47