Single Client Performance - CIFS & NFS on Linux

We have recently revamped our Linux-client testing for NAS units, shifting from IOMeter to IOZone. A CentOS 6.2 virtual machine was used to evaluate NFS and CIFS performance of the NAS when accessed from a Linux client. In order to standardize the testing across multiple NAS units, the following parameters were used to mount the NFS and Samba shares:

mount -t nfs NAS_IP:/PATH_TO_NFS_SHARE /PATH_TO_LOCAL_MOUNT_FOLDER

mount -t cifs //NAS_IP/PATH_TO_SMB_SHARE /PATH_TO_LOCAL_MOUNT_FOLDER

Note that these are slightly different from what we used to run in our previous NAS reviews. The following IOZone command was used to benchmark the shares:

IOZone -aczR -g 2097152 -U /PATH_TO_LOCAL_CIFS_MOUNT -f /PATH_TO_LOCAL_CIFS_MOUNT/testfile -b <NAS_NAME>_CIFS_EXCEL_BIN.xls > <NAS_NAME>_CIFS_CSV.csv

IOZone provides benchmark numbers for a multitude of access scenarios with varying file sizes and record lengths. Some of these are very susceptible to caching effects on the client side. This is evident in some of the graphs in the gallery below.

Readers interested in the hard numbers can refer to the CSV program output here. These numbers will gain relevance as we benchmark more NAS units with similar configuration.

The NFS share was also benchmarked in a similar manner with the following command:

IOZone -aczR -g 2097152 -U /nfs_test_mount/ -f /nfs_test_mount/testfile -b <NAS_NAME>_NFS_EXCEL_BIN.xls > <NAS_NAME>_NFS_CSV.csv

The IOZone CSV output can be found here for those interested in the exact numbers.

A summary of the bandwidth numbers for various tests averaged across all file and record sizes is provided in the table below. As noted previously, some of these numbers are skewed by caching effects. A reference to the actual CSV outputs linked above make the entries affected by this effect obvious.

Asustor AS-304T - Linux Client Performance (MBps)
IOZone Test CIFS NFS
Init Write 55 44
Re-Write 61 46
Read 23 95
Re-Read 23 96
Random Read 13 38
Random Write 45 42
Backward Read 13 31
Record Re-Write 35 710*
Stride Read 21 68
File Write 61 45
File Re-Write 58 44
File Read 16 66
File Re-Read 16 66
*: Performance number skewed by caching effect

 

Single Client Performance - CIFS & iSCSI On Windows Multi-Client Performance - CIFS
Comments Locked

34 Comments

View All Comments

  • jason42 - Wednesday, March 26, 2014 - link

    Anyone know of any small mini-ITX cases that allow for hotswapping hard drives and doesn't look cheap? I'd like to make my one home NAS/media transcoder/HTPC.

    Thanks!
  • Aikouka - Thursday, March 27, 2014 - link

    I don't know if you need this many hotswappable bays, but there's the Silverstone DS380:
    http://www.silverstonetek.com/product.php?pid=452
  • ganeshts - Thursday, March 27, 2014 - link

    Yes, the DS380 is pretty awesome-looking :) It keeps components cool.. only problem is footprint. The U-NAS NSC800 is pretty good too, smaller footprint - same as the DS1812+, but comes at the cost of airflow and cooling capability, obviously.
  • signorRossi - Thursday, March 27, 2014 - link

    Lian-Li has various models, the PC-Q25 offers 7! 3.5" hard drive bays, with 5 of them hot-swap slots.
  • buchhla - Friday, March 28, 2014 - link

    http://www.u-nas.com/xcart/product.php?productid=1... This is the 8 bay version, but they also make a 2 and 4.
  • manmax - Wednesday, March 26, 2014 - link

    It looks like you're using the default mount options for NFS and CIFS mounts on a CentOS 6.2 VM. It would be nice if you actually show what the mount options CentOS uses. For example, using mount without passing any options via the -o parameter to mount a CIFS share could result in the following default options:
    rw,relatime,vers=1.0,cache=strict,uid=1000,forceuid,gid=1000,forcegid,addr=123.123.123.123,file_mode=0666,dir_mode=0777,rsize=61440,wsize=65536,actimeo=1

    rsize and wsize in particular could have a noticeable affect on performance.
  • iwod - Wednesday, March 26, 2014 - link

    Let me get this one thing straight.
    Asustor wins Hands down in terms of performance compared to Synology or Qnap. So if performance is a concern you should get it. Purely because they are using much better hardware.

    However DONT expect its features and software work anywhere as well as Qnap or Synology. If you are after those features, dont get the ASUSTOR yet. Its software properly still needs some time ( a year ? ) to mature. I have heard they are working on it. But as far as i know it still isn't quite there yet.
  • larkhon - Thursday, March 27, 2014 - link

    I think this is even more relevant when talking about the entry level NAS from those brands, Asustor is doing a good job there. But does this one compare to DS414 in terms of performance? price-wise it's the same but it's saturating a single GbE link in many situations hence the second link...
  • beginner99 - Thursday, March 27, 2014 - link

    Why would you even use RAID in these? Look at the rebuild times (assuming it doesn't fail which however is pretty likes with 2 TB+ drives). It seems easier and faster to just copy back the data from your backup.
  • fteoath64 - Thursday, March 27, 2014 - link

    You have a great point on those long rebuild times. It is during such stressing times that another drive might fail so with all eggs being put on the NAS is not such a good idea. Still people will backup periodically to a 4TB or bigger drive just to preserve some of their important data. It is a practise that cannot be forgotten even with a NAS. It STILL needs backup!.

Log in

Don't have an account? Sign up now