Re: ld vs. dkctl in netbsd-8

Louis Guillaume <[email protected]>
Newsgroups gmane.os.netbsd.ports.i386
Message-ID <[email protected]>
On 11/29/18 4:05 AM, David Brownlee wrote:
> On Thu, 29 Nov 2018 at 09:00, Louis Guillaume <[email protected]> wrote:
>>
>> I just noticed this message in the console after a reboot:
>>
>>     raid2: get cache for component /dev/dk1 failed
>>
>> which is a bit troubling, considering I didn't get a similar message for
>> dk0, which is the other half of that raid1 pair (on raidframe).
>>
>> These drives are on separate (slightly different) twa controllers, using
>> the ld driver. I recently updated this driver due to MP_SAFE crashes by
>> applying the patch in this PR (the latest patch)...
>>
>>     http://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=53628
>>
>> Here's an excerpt from the dmesg. Is this cache discrepancy a potential
>> performance hit or should I be otherwise concerned?
>>
>> Any thoughts or help would be great.
> 
> As an empirical check I'd be inclined to perform some write tests to a
> filesystem on the RAID and see if the disks report similar time
> utilisation from 'iostat -x 5'.
> Probably just dd'ing /dev/zero to a file would be fine, though you
> could also run a disk test like bonnie++.
> 
> David
> 

Below are some test results from bonnie++. Seems there is some big 
discrepancy on "intelligent" writes, but not as drastic on "rewrites".

dk1 and dk3 are the raidframe components for "raid2". dk8 is a wedge on 
raid2, and it is used as an LVM PV, which contains the filesystem on 
which bonnie is being run.

Thanks for looking!

--
Louis


# iostat -x 20 |awk 'NR==1||/dk[138]/' &
[1] 17096 16275
#

device  read KB/t    r/s   time     MB/s write KB/t    w/s   time     MB/s
dk1         26.43     18   0.03     0.46      30.76     32   0.03     0.96
dk3         27.53     18   0.02     0.48      30.76     32   0.02     0.96
dk8         44.36     22   0.04     0.93      57.55     14   0.04     0.76

# bonnie++ -u louis -d .

Using uid:1000, gid:100.
Writing a byte at a time...

dk1          0.00      0   0.00     0.00      19.21      6   0.00     0.11
dk3          0.00      0   0.00     0.00      19.21      6   0.00     0.11
dk8          0.00      0   0.02     0.00      27.45      3   0.02     0.08

dk1          0.00      0   0.00     0.00      23.54      4   0.00     0.09
dk3          0.00      0   0.00     0.00      23.54      4   0.00     0.09
dk8          0.00      0   0.00     0.00      35.08      3   0.00     0.09

done
Writing intelligently...

dk1         20.00      0   0.25     0.00      31.93   1142   0.25    35.60
dk3          0.00      0   0.46     0.00      31.93   1141   0.46    35.58
dk8         20.00      0   0.62     0.00      63.73    571   0.62    35.55

dk1          0.00      0   0.30     0.00      31.96   1690   0.30    52.75
dk3          0.00      0   0.66     0.00      31.96   1691   0.66    52.77
dk8          0.00      0   0.89     0.00      63.79    848   0.89    52.80

dk1         26.40      0   0.19     0.01      31.92   1202   0.19    37.45
dk3          0.00      0   0.36     0.00      31.92   1201   0.36    37.43
dk8         26.40      0   0.59     0.01      63.63    602   0.59    37.41

dk1         28.89      0   0.36     0.01      31.94   1649   0.36    51.43
dk3         30.86      0   0.47     0.01      31.94   1650   0.47    51.46
dk8         29.75      1   0.87     0.02      63.75    827   0.87    51.47

done
Rewriting...

dk1         32.00    184   0.51     5.75      31.93   1244   0.51    38.78
dk3         32.00    192   0.55     6.00      31.93   1243   0.55    38.77
dk8         63.82    188   0.89    11.75      63.68    624   0.89    38.78

dk1         31.99    430   0.70    13.44      31.94    883   0.70    27.55
dk3         31.99    451   0.74    14.10      31.94    884   0.74    27.56
dk8         63.93    441   0.94    27.53      63.71    443   0.94    27.54

dk1         31.97    429   0.74    13.40      31.94    887   0.74    27.66
dk3         31.99    455   0.66    14.22      31.94    886   0.66    27.65
dk8         63.97    442   0.95    27.62      63.75    444   0.95    27.66

dk1         32.00    435   0.77    13.60      31.88    906   0.77    28.20
dk3         32.00    464   0.65    14.50      31.88    906   0.65    28.20
dk8         63.97    450   0.96    28.11      63.52    454   0.96    28.18

dk1         32.00    457   0.73    14.28      31.94    934   0.73    29.15
dk3         32.00    475   0.72    14.84      31.94    934   0.72    29.14
dk8         63.97    466   0.95    29.12      63.71    469   0.95    29.15

dk1         32.00    407   0.72    12.73      31.98    852   0.72    26.62
dk3         32.00    444   0.72    13.88      31.98    852   0.72    26.62
dk8         63.96    426   0.94    26.61      63.86    427   0.94    26.62

dk1         31.97    407   0.72    12.70      31.89    829   0.72    25.83
dk3         31.98    419   0.67    13.10      31.89    829   0.67    25.82
dk8         63.83    414   0.96    25.80      63.52    416   0.96    25.80

dk1         32.00    390   0.67    12.20      31.92    801   0.67    24.96
dk3         32.00    407   0.71    12.73      31.92    801   0.71    24.96
dk8         63.95    399   0.96    24.93      63.62    402   0.96    24.96

done
Reading a byte at a time...

done
Reading intelligently...

dk1         32.00    645   0.38    20.14      31.81    105   0.38     3.25
dk3         32.00    682   0.37    21.33      31.81    105   0.37     3.26
dk8         63.41    669   0.49    41.46      63.12     53   0.49     3.26

dk1         32.00   1247   0.72    38.97      11.00      0   0.72     0.00
dk3         32.00   1611   0.62    50.35      11.00      0   0.62     0.00
dk8         63.30   1445   0.86    89.34      13.10      0   0.86     0.00

done
start 'em...

dk1         32.00   1183   0.62    36.97      14.39      0   0.62     0.01
dk3         32.00   1283   0.57    40.08      14.39      0   0.57     0.01
dk8         63.21   1248   0.75    77.04      16.19      0   0.75     0.01

dk1         31.94     64   0.86     2.00      15.42      4   0.86     0.07
dk3         31.87     67   0.80     2.09      15.42      4   0.80     0.07
dk8         31.94    131   0.98     4.08      16.90      3   0.98     0.05

done...done...done...done...done...
Create files in sequential order...

dk1         31.92     64   0.86     1.99      30.27     61   0.86     1.80
dk3         31.82     68   0.80     2.11      30.27     61   0.80     1.80
dk8         31.99    131   0.97     4.09      31.05     58   0.97     1.76

done.
Stat files in sequential order...done.
Delete files in sequential order...done.
Create files in random order...

done.
Stat files in random order...

done.
Delete files in random order...

done.
Version  1.97       ------Sequential Output------ --Sequential Input- 
--Random-
Concurrency   1     -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- 
--Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP 
/sec %CP
xxxxxxxxxxxxxxxx 4G    70  96 49245  50 27616  28   334  99 93338  50 
225.0  32
Latency               763ms    1039ms     351ms   40907us     129ms 
481ms
Version  1.97       ------Sequential Create------ --------Random 
Create--------
xxxxxxxxxxxxxxxx    -Create-- --Read--- -Delete-- -Create-- --Read--- 
-Delete--
               files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP 
/sec %CP
                  16  2032  52 +++++ +++  9557  84  3628  87  5041 100 
5744  91
Latency             62918us      49us    7645us    1606us     463us 
4138us
1.97,1.97,xxxxxxxxx,1,1543487070,4G,,70,96,49245,50,27616,28,334,99,93338,50,225.0,32,16,,,,,2032,52,+++++,+++,9557,84,3628,87,5041,100,5744,91,763ms,1039ms,351ms,40907us,129ms,481ms,62918us,49us,7645us,1606us,463us,4138us
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.