Hello,
My setup has a few ESXi nodes installed on HP DL380 G9, which have Intel 82599 10 Gigabit Dual Port for iSCSI to connect to a 3PAR 8200.
The Intel driver is lastest: ixgbe version 4.5.1
The 3PAR firmware is: 3.2.2 (MU4)
Each vmk adapter has their own DSwitch, Portgroup and physical NIC.
The physical NIC connect to dedicated 10Gbps switch, which have Jumbo frame enable.
DelayACK is disabled.
Recently I got many random event logs about:
Device naa.xxxxx performance has deteriorated. I/O latency increased from average value of 1832 microseconds to 19403 microsecond
Lost access to volume xxxxx (xxxx) due to connectivity issues. Recovery attempt is in progress and outcome will be reported shortly.
After few moment everything back to normal:
Successfully restored access to volume xxxxx (xxxx) following connectivity issues.
Device naa.xxxxx performance has improved. I/O latency reduced from 19403 microseconds to 1201 microseconds.
Checking the performance of physical NIC and storage found nothing abnormal. The throughput is actually low (about 100Mbps)
Could you please share your idea?
Thank you.