跳转到主内容

全闪存FAS (AFF)固态驱动器(SSD)延迟高于预期

Views:
22
Visibility:
Public
Votes:
0
Category:
aff-series
Specialty:
perf<a>AFF</a><a>11000761</a><a>2008083791</a>
Last Updated:

适用场景

  • ONTAP 9
  • Data ONTAP 8

问题描述

  • 在极少数情况下、全闪存FAS (AFF)存储控制器在磁盘层的延迟数可能会高于预期。
  • 在"statit"命令输出或"qos statistics volume latency show"命令输出中可以观察到比预期延迟更大的情况。
例:

"qos statistics volume latency show"命令输出显示磁盘延迟:

cluster::> qos statistics volume latency show
Workload            ID  Latency    Network  Cluster       Data     Disk  Qos Max    Qos Min      NVRAM
--------------- ------ --------   -------- --------   -------- -------- -------- ---------- ----------
-total-                 10.35ms     1.35ms      0ms        0us     9ms      0ms        0ms        0ms
vs1vol0            111  17.23ms        0us      0ms   603.00us  16.63ms      0ms        0ms        0ms
vol1              1234  17.76ms        0ms      0ms   150.00us  17.61ms      0ms        0ms        0ms
vol2               999   4.24ms        0us      0ms   190.00us   4.05ms      0ms        0ms        0ms

 

节点级别命令"statit":

::> node run -node <node> 
node> priv set advanced  
node> statit -b  
node> statit -e  
                       Disk Statistics (per second)  
        ut% is the percent of time the disk was busy.  
        xfers is the number of data-transfer commands issued per second.  
        xfers = ureads + writes + cpreads + greads + gwrites  
        chain is the average number of 4K blocks per command.  
        usecs is the average disk round-trip time per 4K block.  
  disk             ut%  xfers  ureads--chain-usecs writes--chain-usecs cpreads-chain-usecs greads--chain-usecs
/z_nacdot_02_root/plex0/rg0:  
  0b.00.0            5  18.01   15.26   5.11   881   2.04  21.23   112   0.71  14.90   610   0.00   ....     .
  0b.00.2            6  21.06   18.12   4.42   914   2.24  19.31   113   0.69  12.98   769   0.00   ....     .
  0b.00.4           35 950.55  890.12  5.00 2200  40.36  20.00   375   20.07  1.00  6000   0.00   ....     .

  • 在上述输出中、每个4 KB用户读取请求2200 usecs (0.0022秒或2.2毫秒)是所见延迟、即在5个链中、每个传输11毫秒。
  • 依次使用命令"statistics start -object disk -counter io_pending|io_pending_util|io_queued -sample-id disk_queue"和命令"statistics show-object disk -sample-id disk_queue"、您可能会看到以下输出值:

io_pending:1.51
io_pending_util:1.53
io_queued:1.66

  • 在上述输出中 io_pendingio_queued 、大于1表示IO处于待定状态或IO已排队状态、这可能表示固态驱动器(SSD)层出现瓶颈。

 

Sign in to view the entire content of this KB article.

New to NetApp?

Learn more about our award-winning Support

Scan to view the article on your device