SpaceInvaderOne

Thankyou dlandon for that excellent info. I have isolcpu 8 cores for my vms and left the rest for unraid and dockers and pinned emulator tasks to 2 of those cores. Wow i get an extra 1000 score running geek bench3 on my osx vm. And i get no latency on my windows 10 vm. What a difference..... great info thanks so much for sharing

So can they be pinned to dockers or only vms after isolocpus I don't know. I've not tried. I emailed the limetech guys about this and thought i would post what they replied in case others are interested. isolcpus prevents the host OS from assigning the logical CPUs specified in the parameter to any processes (kernel or user space) at boot. Users can manually assign processes back to those CPUs using Linux's cpuset capabilities. Docker and virtual machines both support this. So you can use isolcpus to reserve a set of logical CPUs and then assign docker containers to some and VMs to others and leave the unisolated CPUs for host-based tasks, if you so desire. There may be some advantages to this, there may be some disadvantages to this. With respect to adding isolcpus capability post-boot, so that you can toggle its usage on and off, that is an interesting concept that we have been investigating, but it's not something that's built into Linux's capability set today just yet. We could manually manipulate some things in the user space for this, but it's not something we're tackling as a project right now.

it took all 12 cores 24 threads to get that score. If you want me to benchmark other core counts let me know. My base core speed is 2.4 with all core turbo of 2.7 I use it often but dont run more than 4 cores usually. Just wanted to see what benchmark i could get with all my cores assigned

just read about the memory -m switch to allocate memory as well. Do you think this is worth doing at all. I have never had a docker seem to use much ram anyway?

Ok thanks will do

So i could pin 4 cores to all my dockers, then prioritize which docker has priority, but only run the dockers in those cores, leaving my other cores to pin to vms? eg --cpuset-cpus=7,8,9,10 -c 724 --cpu 7 which would pin 4 vcpus to that docker the give 724 out of 1024 (the defalt docker share) of cpu 7 but normal on others. so then next docker could be --cpuset-cpus=7,8,9,10 -c 300 --cpu 7 so the fist docker gets 75 % of cpu 7, then 8,9,10 and next docker would be 25% cpu7 then 8,9,10. Am I correct in this? Just would like to limit my dockers to 4 cores but give emby higher priority in those cores

ah thanks. yes it was sickrage. I added a path to share temp in docker as i wanted to import some shows into it. totally forgot i did that. thanks

i have a share called temp on my server. Each time i delete it it returns. I have deleted it from gui and also deleted it from command line on all disks. But on reboot of server it always returns. Is it a system generated share thats new in 6.2 beta?

i would be very interested in seeing this aswell

So can they be pinned to dockers or only vms after isolocpus

Thats interesting. I did some testing a few days ago pinning hyperthreaded pairs then pinning none hyperthreaded pairs and running passmark cpu benchmark software afterwards this is my results using a 12 core xeon @ 2.4ghz Using 8 threads or 4 cpu cores firstly 8 threads with paired hyperthreads the passmark score is 7256 next 8 threads all from non paired separate so one thread from different core passmark score is 10417 I as the threads were all on separate cores the speed was faster but i guess wouldnt have been if the other thread on each core was being used by another process. I know this is a bit off topic but thought it interesting as i never expected this result

Yes my chip is fine sold as a 2690v3 but i would put it closer to the 2680v3 based on clock speed. Chip is QEYJ ES 2.4 with a mild bclk overclock i get 17000 passmark score. VT-d is working fine and i have all my vms working. gaming is fine cant notice any difference from my 6700 skylake whilst playing. Any other questions please ask.

Yes i think think it is due to cached performance. i have heard that iozone avoids cache so may be better http://www.iozone.org/

Thanks a lot! so the speed is 144MB/s for write and 149MB/s for read? That's actually less than half of it's bare-metal performance @bluepr0 Dont think its 149mb/s for read speeds i see the write speed as that? " Sequential Write (T= 1) : 149.109 MB/s" @sudonim Are these results from a passed through ssd or froma vdisk on an ssd?

That's AWESOME! and how's the SSD speed? I'm interested on using PCIe NVME SSDs and was wondering how's the performance? Thanks! i just passed through an ssd to the vm using <disk type='block' device='disk'> <driver name='qemu' type='raw'/> <source dev='/dev/disk/by-id/ata-ADATA_SP910SS_8E2820002707'/> <target dev='hdd' bus='sata'/> <address type='pci' domain='0x0000' bus='0x02' slot='0x02' function='0x0'/> </disk> using blackmagicdesign disk speed tests my results were all over the place. writes saying 1000mbs to 300mps and reads from 100mbs to 1000. obviously something is wrong here. on my macbppk pro get constant writes of about 400 and reads 450. however writing and reading files to the disk in vm and opening programmes is all very quick. maybe there is another disk bench programme i can use to test thats better?

Yeah i love my unraid setup. My mac vm is the fastest mac i have. My 2013 macbook pro gets a geekbench score of around 11000 whilst my vm gets 26800, so thats very close to the macpro 2013 with xeon 2697 v2 but at a fraction of the cost!!

Do you have any other pcie slots free on your motherboard to move a card to. On a z170 board i had i could only have cards in slot 1 and 2 , not 1 and 3, or 2 and 3 else they would not be separate

Ok, that was just odd... I tried clicking the map before and nothing happened. This time I just kept clicking it for a while, and suddenly it allowed me to chose a different region. Have been a Mac user since 1990 (and mixed Mac and Windows since 2000), and never encountered that behaviour. Thanks anyways. I was starting to wonder if there was something that needed to be done in the XML or something Still kinda new to both unRAID and Linux in general (or the OS X Terminal for that matter). However I'm having lots of fun experimenting with docker containers and VMs maybe the first time you didnt click on the lock first to unlock it to allow changes?

Thanks so much for taking the time out to post your results. I have just received my e5-2690 v3 es sample cpu. Base speed 2.4, all core turbo 2.7 single core 3.0. Dont know why they sell it as a 2690 ES as its clock speed is closer to the 2680 which has a base clock of 2.5 whilst the 2690 is 2.6? Anyway havent really done alot of testing. Cpu passmark i get 17500. Gaming has no noticeable differences with my gtx 970, except load times are slower on startup. Games play fine but 3dmark seems to crash, i guess because it cant identify the cpu correctly? My windows 10 is on unraid as a vm so vt-d is working fine. I am going to overclock the cpu to try and squeeze a bit more out of it. I have heard the multiplier is unlocked on es samples but not sure. If not i will clock the bclk up to 105.

Very happy with the case. With 6 harddrives in case temps range from 31 to 34 using 2 fans in push pull.

Thankyou so much for your help. I found i actually had 2 backups of the usb i made today. One was missing the super.dat and the other one had it. All is back and running....i am sure i will break something else, but then thats half the fun of it, breaking, learning then fixing!!

Ah thankyou Rob. That file was missing for some reason ,so i have found an old backup with all but one of my disks in it. So i can assign it now but i guess it will then erase it and rebuild from parity?

I upgraded my server to new hardware over last couple of days and had everything working fine. All my disks correctly listed and all shares working. However I had some problems booting from the usb stick about one out of 3 times it would not boot from usb. So i backed it up the usb then formatted it then installed a fresh unraid to test my boot problem. I found problem was caused by my second keyboard causing the rig not to boot from usb for some reason. So after i had sorted that problem i went to put the backup of my usb stick back onto the stick. It then booted fine and i entered my license key. However none of my disks are assigned anymore except the cache drive.(all my plugins are there) I am not sure which drive was parity etc so am worried i will loose all my data if i make a mistake here. I could take server apart an look at each disk and its serial number as i labelled each disk as i put it in but i would rather not have to do this. Please can anyone help? I am using 6.2 beta21

Ah thanks . Shh into server and removed the directories. Now i can run the ca appdata backup fine

Yes i have tried to delete those folders (appdata on disk 1,2,3) using krusader but access to deleting that folder is denied. Is there any other way to delete folders from disks not shares?