Synology SSD Cache Setup & Testing

Поделиться
HTML-код
  • Опубликовано: 2 авг 2024
  • This video, Synology 1621+ Cache Setup and Testing, is the fifth in a series of short videos featuring the DS1621+. In the video I conduct a duplication process prior to and after the creation of the SSD cache.
    0:00 Introduction
    0:27 Setting the Stage For The Test
    1:30 Pre-Cache Test
    2:45 Pre-Cache Test Result
    2:51 Setup of Read/Write SSD Cache
    5:19 Post Cache Test
    6:02 Post Cache Test Result
    6:07 Summary
    7:05 Closing Segment
    7:33 Patron Recognition
    Share this video - • Synology SSD Cache Set...
    Subscribe to this channel - goo.gl/dRbTqD
    Channel Support:
    Buy Tony A Coffee - www.buymeacoffee.com/quiktech
    PayPal - paypal.me/quiktechsolutionsllc
    * Notice Regarding Patreon - I no longer have an active account.
    Earn $100 Cloud Credits from Digital Ocean by signing up at:
    m.do.co/c/41d68f5e11bd
    Amazon Affiliate Storefront:
    www.amazon.com/shop/quiktechs...
    Disclaimers:
    This video is not made for or intended for kids.
    I participate in the Amazon Affiliate/Influencer Program. When purchasing using my Amazon Affiliate links, your price doesn't change. You pay the same, but I do get a small percentage of the sale in commission. This helps the channel to continue creating content you enjoy viewing.
    All images used in this video are my own or taken from pixabay.com.
    About Tony:
    I am a retired educator of 32 years. I started out as an instrumental music teacher and evolved into technology support. After many years of providing technology support to teachers and students, the last ten years of my career, I was in a leadership position of Technology Coordinator and also Supervisor of Technology. My passion has always been helping people. I hope to continue helping people through my RUclips channel.
    Contact Tony at:
    quiktechreview@gmail.com
    Follow me on:
    Twitter @quiktechreview
    Facebook @quiktechsolutionsllc
    Instagram @quiktechsolutionsllc
    Hope you enjoyed this video. Please subscribe, like, & share.
    Regards!!
  • ХоббиХобби

Комментарии • 26

  • @spartan3070
    @spartan3070 2 года назад

    Thanks, Tony amazing video as always !!!

  • @TechMeOut5
    @TechMeOut5 2 года назад

    Great video Tony! Super informative!

    • @QuikTechSolutions
      @QuikTechSolutions  2 года назад

      Thanks Avi. I totally appreciate your feedback. Thanks so much for watching!

  • @CyberGizmo
    @CyberGizmo 2 года назад +4

    Hi Tony enjoyed the video. As I understand it, the Synology SSD Cache works to reduce I/O latemcy. I/O Latency is typically determined by two factors 1) I/O Workload patterns and 2) Storage Media capabilities. SSD Cache typically reduced I/O latency on random I/O which would typically be found in applications using a database for example. In English this means that I/O that occurs across a traditional disk drive (spinning platters and mechanical read/write heads) typically receive the most benefit from SSD Cache because data on the disk platters is forcing the heads to move across the platter and this takes time (adding to the I/O latency). Workloads which involve copies like your benchmark would be accessing the data in more or less a sequential operation (so less head movement). The benchmark times may be getting worse with the SSD Cache on because of the workload type you are using in the benchmark. Since the SSD Cache is unpopulated there will be a Cache "miss" when the SSD attempts to perform the I/O which forces a copy of the data to be made to the SSD, which will of course take extra time to complete. Hope this helps.

    • @QuikTechSolutions
      @QuikTechSolutions  2 года назад +1

      Hi DJ Ware! Thanks for watching and thanks for helping me understand better how the Synology SSD cache works regarding the I/O latency and what may be causing the slow down. I’ve had a bit of a hard time wrapping my head around it. Much appreciated. One of the things I like about this community is being able to learn from one another. Have a great day!

    • @QuikTechSolutions
      @QuikTechSolutions  2 года назад +1

      Hey DJ Ware, I just watched a video from another RUclipsr who did exact same test a few months ago on a DS920+ Intel-based NAS. The SSD Cache actually shortened the process by three minutes. I wonder if the Intel processor is just better suited for this type of task. I have a DS920+ w/20 gig of RAM. I may repeat the test on the 920+ and compare the results.

    • @CyberGizmo
      @CyberGizmo 2 года назад +1

      @@QuikTechSolutions Interesting do you have a link for the video would like to see what they did?

    • @QuikTechSolutions
      @QuikTechSolutions  2 года назад +1

      Yes, here ya go - ruclips.net/video/D9wfMZkLzmM/видео.html. He did a linear I/O a bunch in parallel just like I did. He also included the comparison to Sata SSD as well. But none the less, he got improved results. Also, he is using the DS920+ with the Intel processor.

    • @CyberGizmo
      @CyberGizmo 2 года назад +1

      @@QuikTechSolutions Thanks Tony I looked at his video and to be honest I am a bit horrified at his tests. I have been benchmarking systems for over 40 years and there is process we follow when documenting what we did so others can reproduce the results within a standard margin of error (meaning acceptable differences). He alludes to an off camera process to warm up the data caches but does not cover the steps he did to perform that task. Typically when we would publish results from one benchmark vs is another everything performed during the test to pre-pare, execute and documents the times are recorded. Also the time mechanism he used would be unacceptable in most circles for performance measurement. I learned that lesson early on in my career when i used a stop watch to time the results when i did my first benchmark.

  • @Simonkenteriksson
    @Simonkenteriksson Год назад +1

    Great video :) I subscribed!

  • @raymondjiii
    @raymondjiii 8 месяцев назад

    Great video! Thank you for showing how to create the SSD cache; that was not 100% clear (by doing it myself) what options to choose! Kind of odd that I got a warning about the RAID1 (I use RAID6 on my unit) when that is only choice you can make when setting up the SSD cache.. I was debating on Read vs Read/Write. I honestly, was skeptical of whether I should get this card or not but....I decided to give it a try. I got the E10M20-T1 with the 10GB network port. I do not have any network bridges/switches that run at that speed so I doubt I will use it (plus the DS1819+ has 4 network jacks). I put in two M.2 22110 800GB (All Synology brand parts). I was very surprised to see the results of your test. I wonder if some other tests could be done, especially mounting a drive to your OS and trying to run some 'cp -r....' tests. The results of your test remind where one time I added more memory to a linux machine, and although I was doing very specific kernel testing, I was expecting to see better results (I would have bet the farm that I was going to see an improvement) but instead...I got worse performance. There are so many "knobs" inside any computer system that trying to figure out the test results can be extremely difficult. I bought the DX517 expansion box too. I have not received it yet but it sounds like that has some under sized fans. Lastly, I would LOVE to buy Synology HD's but...WOW .are they really worth the price? That's a huge step up! I was using all Seagate Iron Wolf Pro's but I started switching to Seagate EXOS when I had one drive failure recently. Let me think about what tests I can run. Also, there is some type of analysis tool inside DSM for the cache. I think it's about sizing of it. I have not tried that yet.

  • @keithbennett5179
    @keithbennett5179 2 года назад

    Good video Tony! I was surprised as well. Perhaps the new drives have their own faster cache. That’s my only guess…

  • @cmaciasochoa
    @cmaciasochoa Год назад

    Hello,
    I just got an DS1522+, 20TB HDD storage and 2 ssd 250gb each one for cache, I was wondering if i should first do the setup only with the HDD and then add the ssd or if i can install and do the setup with all the components from the begining and the nas will automatically detect all disk and let me do the job on the very first run?
    Thanks!

    • @QuikTechSolutions
      @QuikTechSolutions  Год назад

      Either way would work. Do whatever is most comfortable for you. Thanks for watching.

  • @raymondjiii
    @raymondjiii 7 месяцев назад

    I can definitely say that on a Mac, if you use iTerm (or iTerm2) and cd to /Volumes/ and then do command line operations. It works order of magnitudes times faster. I tried to "gunzip" a very large file on my Syn nas by double clicking it....I saw no end when it would complete. But using the command line and running gunzip directly it took a reasonable amount of time. Same with running: tar -xvf ... I am running that now and it's about 1 file extracted per second....which I do not consider good but at least things are moving along.

    • @QuikTechSolutions
      @QuikTechSolutions  7 месяцев назад

      Thank you for sharing your experience. Have a great day!

  • @RenoirB
    @RenoirB Год назад

    An alternative use of SSD would be to make a volume, on which we run Docker Compose flight of ElasticSearch/OpenSearch containers with each node writing on then SSD volume folder (e.g. es1; es2; where /volume2/ssdelastic/es1) and each node also with an IP on macvlan.
    Maybe useless to have 2 elastic search node on the same. Maybe have another node, say a RockPro64 or RaspberryPi on same VLAN, and also share a neighbouring macvlan setup.
    All of which using docker compose files

  • @360Link
    @360Link 2 года назад +1

    Normally slower, the first time a NAS device needs to write data to both disk and SSD, if you repeat the test after the first copy the result would be better than without SSD cache.

    • @QuikTechSolutions
      @QuikTechSolutions  2 года назад

      Hi Ivan, thanks for watching & commenting. I thought so too. However, I ran the test off camera several times. The results were pretty much the same. Have a great day!

  • @SethParr
    @SethParr Год назад

    Forgive the ignorance, but why would the SSD Cache apply to a volume and not to a Pool? If you have multiple volumes and the system uses 2 SSD's for one volume, then this seems like a waste as you're giving bias to one volume only.