A customer has iSCSI servers that occasionally perform poorly. The performance issues on a particular
server occur during temporary traffic spikes to that server. Logs show that the number of dropped packets
increases during the traffic spike. In normal traffic conditions, performance is acceptable.
What are iSCSI best practices to fix this problem?
make sure that switches used for iSCSI are non-blocking switches
disable storm control
enable jumbo frames end-to-end
enable bi-directional flow control