GIgabyte Server Activity Corner -- Proxmox, Docker and Config Notes!

@wendell What are the AICs are you using to mount your u.2 drive?

I’ve got a couple of the Startech ones and they’ve been good but (edit very) occasionally flaky. My experience is be that they wouldn’t be recognized during boot and then not show up until a reboot where everything would be fine. Thoughts?

Wendell,
I loved the review on the Gigabyte G-242. I prefer that form factor. I went to scope out prices and wow. Pricey for this one. Is there a similar form factor that you know of in a 2U that could do a good job housing a TrueNas box? It’s hard for me to choose hardware as I do less of that and more software trouble-shooting and db dev and admin’ing. Plus I come from an Apple path no MS and actually looking to move more towards a linux system for myself. So basically I am hardware challenged. I watched that vid you did where you picked up some retired enterprise drive bay and thought “Wow, what a great idea, too bad I don’t know how to find half of that stuff on ebay.” I have no idea what I am looking for.

@wendell

Shouldn’t this be:
systemctl enable enable-turbo
?

5 Likes

Boost and Performance seems to be enabled by default on my 2700.

lscpu

root:~# lscpu
Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
Address sizes:       43 bits physical, 48 bits virtual
CPU(s):              16
On-line CPU(s) list: 0-15
Thread(s) per core:  2
Core(s) per socket:  8
Socket(s):           1
NUMA node(s):        1
Vendor ID:           AuthenticAMD
CPU family:          23
Model:               8
Model name:          AMD Ryzen 7 2700 Eight-Core Processor
Stepping:            2
CPU MHz:             2399.280
CPU max MHz:         3200.0000
CPU min MHz:         1550.0000
BogoMIPS:            6387.26
Virtualization:      AMD-V
L1d cache:           32K
L1i cache:           64K
L2 cache:            512K
L3 cache:            8192K
NUMA node0 CPU(s):   0-15
Flags:               fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb hw_pstate sme ssbd sev ibpb vmmcall fsgsbase bmi1 avx2 smep bmi2 rdseed adx smap clflushopt sha_ni xsaveopt xsavec xgetbv1 xsaves clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif overflow_recov succor smca

boost

root:~# cat /sys/devices/system/cpu/cpufreq/boost 
1

cpupower

root:~# cpupower -c 0 frequency-info
analyzing CPU 0:
  driver: acpi-cpufreq
  CPUs which run at the same hardware frequency: 0
  CPUs which need to have their frequency coordinated by software: 0
  maximum transition latency:  Cannot determine or is not supported.
  hardware limits: 1.55 GHz - 3.20 GHz
  available frequency steps:  3.20 GHz, 2.80 GHz, 1.55 GHz
  available cpufreq governors: conservative ondemand userspace powersave performance schedutil
  current policy: frequency should be within 1.55 GHz and 3.20 GHz.
                  The governor "performance" may decide which speed to use
                  within this range.
  current CPU frequency: 3.20 GHz (asserted by call to hardware)
  boost state support:
    Supported: yes
    Active: yes
    Boost States: 0
    Total States: 3
    Pstate-P0:  3200MHz
    Pstate-P1:  2800MHz
    Pstate-P2:  1550MHz

IIRC For Intel you’ll want:

cat /sys/devices/system/cpu/intel_pstate/no_turbo

Should be 0 for Turbo Boost enabled (double negative - 0 = no_turbo off)

2 Likes

Thank you that was helpful since I’ve got Intel server with Intel Xeon E-2236

I watched the YouTube video. What do you mean when you say that Apple users have unrealistic expectations?

It has been my experience that when safari or some other apple product misbehaves as a result of poor choices by apple that it is anyone but apple’s fault.

Recently, for example, had a customer throw a tantrum because when rotating their iphone, and back, it would add extra margin to the web page. This was, of course, a bug in the browser. It added extra margin for the notch, then forgot to take it away.
That actually had an issue on the browser bug tracker. And the explanation of “yeah, bugs happen. Doesn’t make sense to spend engineering time trying to work around something that will ultimately be fixed without us doing anything in this ecosystem anyway” but they still wanted to be mad about something. I was fortunate I could dig up an issue tracker for that as that isn’t always the case and sometimes work-arounds are warranted.

They never understand that Safari is kind of becoming the IE6 of web browsers. Browsers have bugs.

Also, apple’s app development policies and the arbitaryness. “When will my app publish?” "idk, whenever apple blesses it. could be a day. could be a month! who knows! but expect to have to change the icons 5 pixels or change the shade of gray from #cccto #bbb or some other arbitary thing. That apple could be so random or arbitrary seems incredulous…

4 Likes

What about huge pages?

Can you provide some more information on how you used this server to automate mobile device testing? Could you go into more detail on the software side of things?

Soon™. Start with the basics. Have you used chrome/ chromium headless? It s got a ton of cool features

chrome-launcher/chrome-flags-for-tools.md at master · GoogleChrome/chrome-launcher · GitHub

I’m interested in the reasoning for choosing Proxmox over something like vmWare for this?

I’m working on a project myself, for a Machine Learning -type system using a Dell R840 with 4x Platinum 8260, and have installed Proxmox because of the design requirements.

You mentioned both in the Video on Level1Linux and this post running Docker in LXC or a VM. I’m starting now with my own homelab server after renting a VPS for the last 1.5 years. I’ve planned to setup docker in a VM, but now it seems like a LXC container may be the better choice. I’ll have services accessible from the internet (probably piped through my vps), but don’t know enough about the security risc using containers (both LXC and docker). A VM seems to be the safest choice, but has the most overhead (altough I’ve got pretty good experience with KVM and PCI passthrough before)

I think if you go to the trouble of log-forwarding your docker container, and keep an eye on it, the security aspect isn’t too bad. Everything is a trade-off. Especially if you have ingress filtering farther up the chain.

1 Like

Do you think you should add a step to set the max ARC size so thatZFS doesn’t try to eat all the RAM since the server is mostly going to host VM’s and containers.

/etc/modprobe.d/zfs.conf

# https://pve.proxmox.com/wiki/ZFS:_Tips_and_Tricks
#
# don't let zfs use more than 8 GiB of ram for ARC
options zfs zfs_arc_max=8589934592
# disabling prefetch is no longer required
options zfs l2arc_noprefetch=0

On my proxmox I setup to use tuned instead with the virtual-host profile.

virtual-host

A profile focused on optimizing performance in Red Hat Enterprise Linux 7 virtualization hosts.

This profile favors performance over power savings by setting intel_pstate and max_perf_pct=100 . It also decreases the swappiness of virtual memory. This profile enables transparent huge pages and writes dirty pages back to disk more frequently. It uses cpupower to set the performance cpufreq governor. It also sets kernel.sched_min_granularity_ns to 10 μs, kernel.sched_wakeup_granularity_ns to 15 μs, kernel.sched_migration_cost to 5 μs, and vm.dirty_ratio to 40 %.

Lastly, I don’t know about Gigabyte servers but for my HPE server I also had to set the Virtualization - Max Performance setting in my UEFI.

2 Likes

Bane of my existance. Like IE, when it worked it was great but when it breaks, expect the user fits. Though I have to say being a Mac Admin I note the fits of Wintel users are much more reasonable than my fellow Mac fans. My theory is that creatives are more emotional and since the latter platform historically has been the choice of a slightly higher number of creatives, responses are more emotional. I can’t say this has made my skin any tougher. :-/

Awesome, I’m looking forward to it. Yes, I have used headless chrome for things like web scraping. It really is a fantastic tool.

It’s pretty incredible how active this community is btw. I’m just getting into all of this (I’m a software developer and CS grad) so I’m still new to this side of things.

Wendell, amazing explain, more words is best!!, don’t hesitate to add a lot of explanation for newbies like me… I bookmarked this page, please could yo change to…

systemctl enable enable-turbo.service

for the future… :slight_smile: D

@wendell , Please share your experience on how to get the best experience out of and s7150 in VDI !! I myself and experimenting with an AMD 7150X2 for VDI with windows 10 vms

Is this still needed with current Proxmox 7.3. Am deploying the same set up here with an EPYC 7402 and want to set it right and forget about it.