高写入延迟， MetroCluster 上出现背靠背 CP 和 SCSI 命令错误
适用场景
- ONTAP 9
- MetroCluster FC
- Brocade 后端交换机
问题描述
-
客户端报告延迟较长
-
scsi.cmd事件日志中报告的错误：
示例：
[NodeA: isp2400_intrd: scsi.cmd.abortedByHost:error]: Disk device ATTO:8.126L1504: Command aborted by host adapter: HA status 0x4: cdb 0x9a:0000000121806060:0001:01a0.
Fri Apr 16 10:01:02 +0200 [sminfsa01001p: isp2400_intrd: mlm.excessive.TPlatency:notice]: Average latency of 258239us on target port 2100001086ad1f00, has exceeded 250000us
- 低 SFP TX 启动交换机端口：
Switch:admin> sfpshow 8
…
RX Power: -1.5 dBm (708.8uW)
TX Power: -5.6 dBm (273.8 uW)