[Support] binhex - SABnzbd


Recommended Posts

  • 3 weeks later...

Firstly, Binhex, I've been a long-time user of your docker repositories. Just awesome. Thank you. 

 

For what it's wroth, I too am having unusual activity recently with sanznbdvpn, after a few downloads my unpacks become inexplicably slow, which coincides with downloads becoming slow - download speeds start 'yo-yoing'. 

 

Thinking this may have been related to par2cmdline-turbo, I tried Binhex's updated docker image tagged 'test' which includes par2cmdline-turbo. To be fair, normal unpack speeds seem quicker, it seems to work. But I am still getting 'yo-yo' like download speeds which coincide with strangely slow unpacks. These occur after a bit of sustained downloading. At first I thought it was my ISP or VPN (PIA), but this wouldn't explain the slow unpacking.


This was tested using both 1) a RAID1 BTRFS 2x SSD Cache, and 2) an XFS 1x SSD Cache. It doesn't seem to be related to storage write/read speeds.

 

I've also set direct_unpack_threads (Default 3) to 6, to give sabnzbd a bit more grunt - no joy (I'm running a 13500 with 20 threads + 32GB of Ram, which should be plenty). 

 

Here is an example of the yo-yoing:

 

1.thumb.png.11eec4b18ae69444b8743f2a4b69b7c8.png

 

You can see that the speeds have gone from a locked 88MBps (where I cap it, I have a 1Gbps connection), and started yo-yoing. This behaviour coincided with an unpack process slowing down (pictured below). 


Pausing the downloads, letting the unpacks finish and restarting fixes the issue, albeit temporarily:

 

2.thumb.png.db75450db6440d732d5528f94b24a0da.png

 

After a while, the yo-yoing behaviour starts again. Interestingly, the time it takes for yo-yoing to start varies. You can see below that downloads held constant at 88MBps until the strange unpacking/download speed behaviour started again. 

 

3.thumb.png.91c6e705a8dbfdd5ab8b86caaabb6c15.png

 

I thought I'd share my experience as this seems to be impacting more and more people. 

 

My motherboard has a Dragon 2.5G LAN + Intel® Gigabit LAN, the same behaviour occurs on both. 

 

Hope this post helps.

 

EDIT1: I should say I'm new to Unraid (Jan of this year), but have been using Binhex's repositories on Synology NAS systems for years. Also, Unraid team - you are amazing. An absolutely great platform - thanks to you too!

Edited by nasbox
Link to comment
15 hours ago, nasbox said:

Pausing the downloads, letting the unpacks finish and restarting fixes the issue, albeit temporarily:

If the yo-yo speeds happen when unpacking or par running andthe yo-yo stops when you pause the unpack/par then it most probably is IO or CPU related, check your CPU and IO usage next time it happens, you could also try downloading and unpacking to different locations, so download to SSD and unpack directly to the array for instance, or even to RAM if you have enough.

Link to comment
10 hours ago, binhex said:

If the yo-yo speeds happen when unpacking or par running andthe yo-yo stops when you pause the unpack/par then it most probably is IO or CPU related, check your CPU and IO usage next time it happens, you could also try downloading and unpacking to different locations, so download to SSD and unpack directly to the array for instance, or even to RAM if you have enough.

 

Thanks for the prompt response. 

 

Overnight, I decided to test whether FUSE was causing the issue. I remapped /mnt/user/downloads to /mnt/cache_patriot/downloads - no joy. Not FUSE related.

 

After reading your response I remapped downloads & unpacking (.../usenet/incomplete & .../usenet/complete) to RAM (via Unraid's /tmp/... folder). My logic: I wanted to identify whether it was a CPU problem (which I doubted) or IO issue (which, admittedly, I also doubted) - problem solved! Downloading, unpacking and or repairing in RAM resolved the yo-yo problem. 

 

I have to admit I was surprised, my current Cache Pool arrangements is:

 

4.png.e03c7097c1e947e0b032e2c616bac45d.png

 

Cache_nvme: 2x Raid1 NVME: For Docker, VMs, etc

Cache_patriot: 1x Cheap Patriot SSD: For downloads, transcoding, and anything with high IO

Cache_ssd: 2x Raid1 SSD: general Cache for data requiring redundancy

 

The yo-yo occurred on both the patriot & raid1-ssd cache pools. It didn't occur straight away, for between 20m and ~1h of sustained 88MBs downloading, SabNZB was fine downloading and unpacking files. Randomly, it would start, and occur regularly, not going away until a restart. After the restart, the cycle would start again. This pattern of behaviour made me think it was a strange i) Unraid issue, ii) Sabnzb issue, and or iii) a combination of both. This thinking was compounded by the problem having only started occurring in the last ~2-3 months (it ran fine on the cache_ssd pool for a long time, when I first setup Unraid).

 

I'd love to continue downloading to RAM. But my server lives in a Jonsbo N3 which requires an mITX mobo, max RAM at this stage is 64G. This isn't sufficient for large nzb files. 

 

I have moved my downloads folder to the nvme Cache Pool which works fine (just like RAM). This is probably not an ideal long term solution as the high IO is using up 2x NVMEs - the whole point of the Patriot SSD was to avoid this. 

 

I may have to find a good SSD which can deal with the IO.

 

Thanks for your help. Excuse the length of this post, I wanted to share in the hope it helps others. This was driving me crazy!

 

PS. If anyone can recommend an SSD which works well with SabNZB, please shout out. 

 

Link to comment
6 hours ago, nasbox said:

The yo-yo occurred on both the patriot & raid1-ssd cache pools. It didn't occur straight away, for between 20m and ~1h of sustained 88MBs downloading, SabNZB was fine downloading and unpacking files. Randomly, it would start, and occur regularly, not going away until a restart.

I would suspect when the machine is freshly booted then the Patriot SSD has a clean cache (onboard of the SSD) and as you use the SSD more the cache fills up causing read/writes to slow down, I mean it could be a firmware bug for the Patriot SSD, worth investigating if there is an update available, failing that ditch that sucker an get yourself something a little faster - the Samsung Evo series are always solid, I have an Evo 850 that has been going strong as my download drive for literally years!, then i use a Samsung 970 Po for Docker image, VM's and appdata - I cannot guarantee it will work for your specific use case, but it should.

  • Like 1
Link to comment
On 10/20/2023 at 7:19 PM, binhex said:

I would suspect when the machine is freshly booted then the Patriot SSD has a clean cache (onboard of the SSD) and as you use the SSD more the cache fills up causing read/writes to slow down, I mean it could be a firmware bug for the Patriot SSD, worth investigating if there is an update available, failing that ditch that sucker an get yourself something a little faster - the Samsung Evo series are always solid, I have an Evo 850 that has been going strong as my download drive for literally years!, then i use a Samsung 970 Po for Docker image, VM's and appdata - I cannot guarantee it will work for your specific use case, but it should.

 

Thanks for the response - I took your recommendation and bought a Samsung EVO 870 500g HD for my downloads - works perfectly. I didn't realise how different SSDs could be.

 

I also reconsidered how SABnzb downloads, I've set the incomplete directory to the new Evo SSD to minimise IO, and the complete directory to the Cache Pool (where movies go). This minimises IO while allowing Sonarr/Radarr to use Hardlinks when moving files to the media library. 

 

Thanks again for your help, and SSD recommendation - all works perfectly now.

  • Like 1
Link to comment
  • 2 weeks later...

Pretty sure I've tweaked something in my settings somewhere regarding how sabnzbd is writing to the array.  Originally I  was saturating my gigabit internet - now hitting only 85 Mbps.  Have tested on another device with same settings on sabnzbd and it hits the limit.

 

downloads are set to hit Cache

medis pool set to hit Cache then Array.

 

What have I set wrong?

image.png

Link to comment
On 11/2/2023 at 3:24 AM, eb98jdb said:

Pretty sure I've tweaked something in my settings somewhere regarding how sabnzbd is writing to the array.  Originally I  was saturating my gigabit internet - now hitting only 85 Mbps.  Have tested on another device with same settings on sabnzbd and it hits the limit.

 

downloads are set to hit Cache

medis pool set to hit Cache then Array.

 

What have I set wrong?

image.png

 

If you think you're write speeds are slow due to downloading to the Array (ie. IO speeds), check if this is the case by changing the download/incomplete & download/complete folder to RAM. You can do this by mapping these folders to either /shm (a folder which gives you access to 50% of RAM), or /tmp (all of RAM). In doing so, you should make sure that you have ample RAM to download your test file without maxing out RAM (that would be bad). If this downloads at max speed, then it's likely an IO issue. I used this method to prove my slow write speeds were IO.

 

I never download to the Array, for me it's full of HDDs, which are much slower than SSD. Plus, if you have a Parity drive, you're just wasting IO. Setup a Cache drive with an SSD and download to that. I've recently moved to a Samsung EVO, which works nicely for me. 

 

Added point: If you're using the Binhex VPN containers, I have found in the past that the DNS name servers I was using were very slow. Slow DNS servers resulted in slow download speeds. As the DNS servers were intermittently slow, the slow download speeds were also intermitent, making diagnosing this problem hard. I changed the standard name servers to PIA's DNS server (I use PIA as my VPN). If you don't use PIA, you could try Cloudflare's or Google's DNSs.

 

Good Luck. 

Link to comment

My Sab is not downloading anything all of a sudden:
Indexers are working fine (and active) and can grab:
this is a test download of a file which i had already succesfully downloaded previously, just to prove it was available. Some will start but always show missing articles, and thats for anything. It suggests to me the problem is with newshosting.com
image.thumb.png.1db753d7dcc300ac4f44506f6abd4c4c.png

 

NZB server is newshosting.com which is active and tests good:
image.thumb.png.7b518d3b4a675ef2d0bc163183d2f941.png

logs:
image.thumb.png.90dd6e9eec1b31c7d155ea3027e18db4.png

more testing now suggests a sab problem: sab test DL failed
image.thumb.png.f58bda0fb2aeab20859b34264e6ad0e8.png

Edited by K1LLA_KING_KING
Link to comment
  • 4 weeks later...

Has there been any changes to changing the path for downloads?

 

I changed the data path to use /data/mnt/user/downloads/sabnzbd/ but for some reason cannot browse anything inside even with RW permission.

 

Sounds like im missing something simple, using similar method has worked for Sonarr.

Link to comment
  • 2 weeks later...

I'm having an issue with sabnzbd sending SMTP messages to my local mail server (10.13.2.220) I configured VPN_OUTPUT_PORTS=25, which is the port for my SMTP server. But test emails are failing to be sent with the VPN up. 

 

sabnzbdvpn:
    image: binhex/arch-sabnzbdvpn:4.0.3-1-03
    container_name: sabnzbdvpn
    privileged: true
    environment:
      - VPN_ENABLED=yes
      - VPN_PROV=protonvpn
      - VPN_CLIENT=wireguard
      - STRICT_PORT_FORWARD=no
      - ENABLE_PRIVOXY=yes
      - LAN_NETWORK=10.13.2.1/24,10.189.177.0/24
      - NAME_SERVERS=1.1.1.1
      - VPN_INPUT_PORTS=50005
      - VPN_OUTPUT_PORTS=25
      - DEBUG=true
      - UMASK=000
      - PUID=0
      - PGID=0
    volumes:
      - /mnt/data/:/data
      - /apps/docker/sabnzbd/config:/config
      - /etc/localtime:/etc/localtime:ro
    ports:
      - 8080:8080
      - 8090:8090
      - 8119:8118
    restart: unless-stopped

 

Edited by Globe89
Link to comment
  • 4 weeks later...

Can someone help me understand the paths I need to enter to get SabNZB working with Radarr, Sonarr and Plex

I have created a share called media, within that, two folders Films and TV.

I want films ans Tv series automaticaly placed in those folders when the have been downloaded.

Link to comment
5 hours ago, Slin said:

Can someone help me understand the paths I need to enter to get SabNZB working with Radarr, Sonarr and Plex

I have created a share called media, within that, two folders Films and TV.

I want films ans Tv series automaticaly placed in those folders when the have been downloaded.

 

Your download path should be to a cache disk, so: /mnt/cache/Downloads/sonarr/

The Container path should be default: /data

 

Your media paths should be /mnt/Media/TV and /mnt/Media/Films

The container path should also be default: /media

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.