SundarNET

just adding in from my observations around 10 Mbit is not too bad around 20 Mbit gets me 50% load then higher it gets worse ill edit this comment as I see more! 2.3GB file downloading at about 6-10Mbit and im getting 90-100% CPU loading on all 80 cores I had to pause the download it was going to fry my server now CPU loading is at 7-15% ish this is the strangest thing I only had to pause the one download im going to try and start something else and see what goes down! 1.3GB file @ 10-20Mbit and getting about 20-50% loading I just shuffled a new download to the top there is also another smaller file downloaing 149MB file @ 200-500Kbit now 3 files going @ 12-27Mbit 20-40% CPU loading Im going to throw a different big file at it now 11G file @ 10-15Mbit 14-25% CPU loading ok bad test the 11G download is lots of smaller files as opposed to the single 2.3G file off to find a huge single file I need to get ok added a new file 6.4G file @ 7-10Mbit 60-80% CPU loading the CPU loading is actually fluctuating quite a bit from 30 ish up to 80-90% it started to get into the high 80-100% fry my server load so I had to pause my downloads ok so it seems to be something with large files around 2GB or larger approx and download speeds of over 10Mbit

funny you mention that I dont think it is the issue 'I could be wrong' I just upgraded from the internal quad port LAN 2x 1GB 2x 10GB in my Dell R910 server, ethernet connections all running @ 1gb, to a dual port 10G fiber optic card and found it exactly the same

I must agree it was fine for a day but now yup 100% again and its directly correlative to the download speed anything over 10mbit peaks CPU use thank you MammothJerk for that one! and I keep complaining because my system is a quad xeon server with 40 physical cores and 40 HT cores and when you see that shit maxx out its worrying

I have been getting the same how old is your docker I had to remove mine and reconfigure it to stop this issue but just inmagine seeing that on 80 cores

I had to resolve this the only way I know by removing the image and deleting the files 'renamed the appdata folder just in case' now working perfectly no 100% CPU spikes and as I stated initially for it to spike on 80 XEON CPU cores is ridiculous do you want my old deluge config files to see if theres something strange going on

imagine seeing it on 80 cores man its scary! but yes i usually close the container for a bit then restart it and it goes fine for a few days I suspect its some sort of file checking going on maybe?

Im no expert but it sounds like you know what you need to do ... maybe you should just bite the bullet.

Im not fully sure but I think my DelugeVPN is loading my CPU's @ 100% 'I have 4 x 20 core Xeon CPU's so this should be really almost impossible ' I also have Deluge limited to about 8 CPU cores not all of them and it only does it sometimes, top CPU monitoring tells me sha256sum is the process that is creating the load but if I shut down Deluge it stops any suggestions please

https://www.ebay.com.au/itm/LSI-6Gbps-SAS-HBA-LSI-9200-8i-9211-8i-IT-Mode-ZFS-FreeNAS-unRAID-AU-seller/143058624384 you cant use RAID cards happily with URAID OS

try this command sg_format --format --size=512 -v /dev/sg3

So yay !! I'm pretty sure I got it worked out mapping /steamcmd/rust for data/settings retention scroll to bottom click show more settings change Mapping Directory to the location you want to have the Docker container store the files I'm 90% sure the container will create /steamcmd/rust the first time you click apply, Ive been stopping my container and using krusader to move the files from /appdata/Rust-Server to your chosen mapping/save location I also made 2 separate docker containers by renaming one and giving it a separate IP address and a separate /steamcmd/rust2 mapping. that way it keeps it separate but also doubles up on storage use. I have my 2 RUST servers running parallel network is custom : br0 then assign an ip in your network subnet one for each rust server leave all the ports as normal, i didn't have time to find out if you can run serve RUST on different ports and have it all forward nicely ' I however have the luxury of 2 WAN connections so can cheat by forwarding one server to each separate WAN connection. 'please let me know if there are other options.

also to add in Mdarkness1988 I love your rust container its great ive managed to get 2 servers running in parallel just have to get my second modem forwarding ports correctly

just asking again for the thread Mdarkness1988 please as I cant find it so far how to map /steamcmd/rust to the host for data retention ive been trying to use path map in container /steamcmd/rust :container path /mnt/user/Data/ :host path and it just spits an error and makes the container disappear can just add it again with the correct settings and it returns easily what am I doing wrong?

I like just found this and yes i love it to elaborate in terminal cd /boot mkdir extra cd extra wget " paste the sg3_utils download link here they change it a lot " "google search sg3_utils slackware" reboot there's your presto

yeah I've been checking and I'm pretty sure I'm completely wrong and you should always google search your issue to get other options. regarding drives and formatting, i think most drives, SAS or not - need a logical byte sector of 512 then a physical of 4k, and over 2TB needs to be GPT partition map to see all the drive 'for windows' most other systems will format accordingly and I assume UNRAID is the same, let it do its own default format it should make it right and see the full size of the drive, another possibility... assuming your are fresh with UNRAID like me, is your parity drive bigger or the same size as this 4TB as that is the first possible issue that your parity drive is only 3TB? please give more info always to assist proper diagnosis

ok it seems 'I may be wrong, always add as much information as possible' that you may have formatted a normal drive from 4096b sectors to 512b format it back to 4096b block size this tool is for SAS drives mostly and non SAS drives NEED to be 4096byte block size to have their drive size correct once again without all the information I can only speculate that this is your issue

what model brand type of drive?

quick Q is it showing that in windows? if so you may just have it formatted as MBR and not GUID, I could be sadly wrong there though

hmm not nice you will have to do some google searches to find out this issue Ill have a hunt when not busy, I assume it is something to do with a simple command that needs changing

Just to confirm the install process works I just did 1. and 2. on my main UNRAID machine and it installed perfectly and sg_scan -i showed the list of drives, I had no need to format so will be doing that a bit later

Reformat a SAS HDD to different block sizes mainly 512 to use in UNRAID This took me a few hours to find and work out but was so much needed OK so I have just now done this for myself by installing sg3_utils onto my UNRAID OS using installpkg all using terminal 1. download the package into a tmp dir # wget http://slackware.cs.utah.edu/pub/slackware/slackware64-14.1/slackware64/l/sg3_utils-1.36-x86_64-1.txz 2. run this from that tmp dir after the download to install sg3_utils # upgradepkg --install-new sg3_utils-1.36-x86_64-1.txz 3. use this command to show SAS HDD's # sg_scan -i 4. this command to format 'obviously /dev/XXX should be the HDD u wish to format MAKE SURE ITS THE RIGHT ONE! # sg_format --format --size=512 -v /dev/XXX this has been allowing me to reformat the block size and use previously non usable drives saving buttonnes of money WARNING this format will destroy a HDD if interrupted during this process if you can a UPS is recommended have a great day I love UNRAID!