NAS / OpenMediaVault

[fire219]

Well, I tried iozone with 1GB "file" :

Code:

                                                     random    random     bkwd    record    stride
             kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
        1024000       4    21398    30223    77006    76949    74991    28235
        1024000      16    54907    77383   245716   246803   239836    73787
        1024000     512   118921   132870   322434   312990   323876   126180
        1024000    1024   117895   147201   304274   318403   316472   121843
        1024000   16384   125973   120491   320027   321014   320234   129416

Past reclen=16, speed is ok for me (I could live with it ).
So why is speed so low when transferring file from my PC to the SSD on the Rock ?

Those iozone numbers look good, though a little low on the write. 320MB/s read (as is shown on the final test) is close to the maximum of USB3 with overhead. Performance is always going to be terrible on small file chunk sizes (that's the reclen column -- record length) on any system, even with an SSD.

But this shows that it's not a problem with the Rock64 (or at least not with the USB3 controller and UASP drivers). I'm not sure what could be causing the problem. You may want to check network performance on your R64 (try using iperf3).

[Metalazzo]

I tried using 4GB file for iozone :

Code:

                                                            random    random     bkwd    record    stride
             kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
        4096000       4    17511    27478    60249    62087    15937    24649
        4096000      16    54121    70899   133485   133685    45866    67819
        4096000     512   125061   143120   150598   150483   122570    97956
        4096000    1024   127317   152012   169291   161070   148902   106556
        4096000   16384   130990   150823   156137   153770   162362   120992

SSD is a 128GB from sansdisk.

I'm really puzzled.

[fire219]

I tried using 4GB file for iozone :

Code:

                                                            random    random     bkwd    record    stride
             kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
        4096000       4    17511    27478    60249    62087    15937    24649
        4096000      16    54121    70899   133485   133685    45866    67819
        4096000     512   125061   143120   150598   150483   122570    97956
        4096000    1024   127317   152012   169291   161070   148902   106556
        4096000   16384   130990   150823   156137   153770   162362   120992

SSD is a 128GB from sansdisk.

I'm really puzzled.

Did you try running iperf3?

[Metalazzo]

I tried using 4GB file for iozone :

Code:

                                                            random    random     bkwd    record    stride
             kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
        4096000       4    17511    27478    60249    62087    15937    24649
        4096000      16    54121    70899   133485   133685    45866    67819
        4096000     512   125061   143120   150598   150483   122570    97956
        4096000    1024   127317   152012   169291   161070   148902   106556
        4096000   16384   130990   150823   156137   153770   162362   120992

SSD is a 128GB from sansdisk.

I'm really puzzled.

Did you try running iperf3?

I decided to format the disk and change the "format".
Previously, it was in NTFS, I didn't remember the fact that it was an old laptop disk. I was sure that I alreayd tried it with my bananapi (and so, that it was formated in EXT4).
I formated it in BTRFS and now, speed are at last, OK (no more limit at 25MB/s like the result with a reclen of 4)
40-50MB/s with an old 500GB 2"5 disk, I need to test again the Sandisk SSD. I'll do it later.

[stuartiannaylor]

I tried using 4GB file for iozone :

Code:

                                                            random    random     bkwd    record    stride
             kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
        4096000       4    17511    27478    60249    62087    15937    24649
        4096000      16    54121    70899   133485   133685    45866    67819
        4096000     512   125061   143120   150598   150483   122570    97956
        4096000    1024   127317   152012   169291   161070   148902   106556
        4096000   16384   130990   150823   156137   153770   162362   120992

SSD is a 128GB from sansdisk.

I'm really puzzled.

Did you try running iperf3?

I decided to format the disk and change the "format".
Previously, it was in NTFS, I didn't remember the fact that it was an old laptop disk. I was sure that I alreayd tried it with my bananapi (and so, that it was formated in EXT4).
I formated it in BTRFS and now, speed are at last, OK (no more limit at 25MB/s like the result with a reclen of 4)
40-50MB/s with an old 500GB 2"5 disk, I need to test again the Sandisk SSD. I'll do it later.

You did right the info might be on the Armbian forum but with NTFS how its mounted and the tune2fs parameters make a huge difference.
I can not remember the exact details but a google NTFS-3G tune2fs should help.
Some guru should post a test specification and setup script so we are all singing from the same hymn sheet.
http://www.cnx-software.com/2017/08/07/r...enchmarks/

Samsung EVO 64GB with data=writeback option set: sudo tune2fs -o journal_data_writeback /dev/mmcblk0p2
https://forum.pine64.org/showthread.php?...&pid=12700

Sure someone will supply the optimal settings

[Metalazzo]

I tried another file transfer and the speed for a 2GB file was at 100MB+/s.
When trying iozone, the result I had for 4k reclen test was similar to the speed I had when trying to transfer file.
So my guess was that the system file was not the right one... Now, I'll be able to go to sleep earlier tonight.
Being a noob sure is tiring ^^;

[Metalazzo]

Hello guys,
The noob is back.
I have a question : how can I update my installed OMV with the latest version of Sir Ayufan ?
I mean, I have version 0.4.17 and I see a version 0.5.1.
The following command :
sudo apt-get update -y
sudo apt-get install linux-rock64 -y
does not upgra anything...

Am I obliged to install all the *.deb available here (https://github.com/ayufan-rock64/linux-build/releases) ?

Thanks for your help

[ayufan]

Don't update. Once I mark it as a release (when it is tested and stable) you will be able to do it.

[alexcismaru86]

hello

please help me with this

I have installed omv on a microsd card, fired it up, configs, everything great until I restarted the board to see if the ip address remains the same.
well, guess what? not only the ip address but even the mac address changes every time I reset the board or unplug the power supply.
how can mac address change? isn't every device hardcoded to generate the same address every reboot?

so of course my router assigns different ip every time cause of the mac.

what can I do?

rock64 2g ram 16g microSD Sandisk Ultra class 10, asus nt56u diamond

thank you!

[alexcismaru86]

hi

I think I found the problem
os cannot read the mac address from the device, that's why generates a new one every time
anyone have an idea to solve this?

"rk_vendor_read eth mac address failed (0)"




thank you!