#the-water-cooler
1 messages ยท Page 59 of 1
tech related ๐ another project
lol
Fancy igus chains tho
Thumb ring man says what
Now I should show you my prince Albert
M o d s
yeah, the camper is the temporary home. building a new one for it
Literally?
I say this unironically way too often
i'm 43, do i count as an adult? ๐
Definitely not related to pegging
I'm 39, I wouldn't count myself as an adult
Youโre here so no
LMAO
I thought you wore big boy pants
Iโm not wearing pants right now
Texas summers be like
Mass shootings tend to leave you pantless
Thoughts and prayers
If only they helped
๐ญ
none up here in Maine yet, and i hope it stays that way
I'm disappointed there is no Secondarie
One heck of a Release Party that was
Sir this is a shitpost den
Maybe we need general shitpost and ha related shitpost chan.
There is always room for more shitposting. Especially when most channels don't allow images
A succulent Chinese meal?!
Thatโs just general with extra steps
They are happy with the state of the server and there is no suggestion box this what we get lol
I tried getting us embed images / pictures in #hardware-archived and failed
Mods know best
It must be that way
Users are also idiots who post text as images
Weโll hardware has a very good use case
Yes
One, embedding images to links.. of hardware
Two, sharing images.. ofโฆ hardware
The wear was hard
Itโs also just a core tenant of discord and very much confusing for people used to every other discord ever
puts on his MLK hat
I have a dream
Pictures
I'm so glad I haven't worked with users for decades now
Thought of you @steel crag
Isnโt Cisco slowly dying
We can only hope
And then there's the faction here that treats the frontend like some giant bug ๐
where yaml? ok on GH ๐
I love the network tab, going to try to recreate something similar
Network tab ๐ค
go watch the party
The network, but with the graph from the energy tab
that is also impressive yes
Yea, I need more smart sockets so I know where my power is going
you customize your dashboards?
My main dashboard
Which I mostly use because CEC is broken in my setup
Why is everything but deckenlampe and temperatur in english
Because I'm the only user, and don't really care
Pretty sure my room names are out of 3 languages
What do you think about this design?
https://www.figma.com/proto/NZAIUoXeYOjYBRk6EG2LB1/Home-Assistant?type=design&node-id=54295-34936&t=8RTxMzKL6cOUD4D6-1&scaling=min-zoom&page-id=54295%3A34798&mode=design
if you want feedback on your UI then #frontend-archived is the channel you wanted
Figma? is that a fork of Ligma?
heh up to 10 years
Who the hell is Steve Jobs?
up to 10 years"
idk my interpretation is that it means if the power goes out the lights must restore properly (except if the power goes out for >10 years)
I read it the same way
I have a Proxmox installed on my home server and Home assistant is running on a VM, currently most of my containers (tailscale, adguard, uptime kuma, etc.) are installed as an add-on on my Home Assistant instance. And I have another VM running open media vault which also has some containers like portainer and plex. It is all working well but as you can see it is a mess.
So what is the best way to install containers in my situation?
Run a Debian VM
Personally though, I'd blow it all away. Run Debian on the host directly, use KVM for HAOS and run everything else in Dockerr
why HAOS over container in that setup? ๐ค
I'd imagine less worries about dependencies
what about adding a new VM with ubuntu server os, and installing docker on it and use portainer to install other containers?
that's the fancier version of running a debian VM
though I have no idea how and if portainer works ๐ฎ
would debian be better?
I mean... I wouldn't, but some people actually like HAOS
I get that it's cool to just run it as appliance. But I don't see why you'd run it i you already got a container env
then again, /me is probably one of the few people running HA in a restricted k8s environment
i like locking down the application's environment, while i have full privileges on the orchestration ๐คท
is there a way to "export" containers from HAOS ?
#installation-archived could help you plan a migration
huh, I still run HA as root in the container. i should figure out why at some point
we should tone the haos discrimination down a bit, they barely make it down to this channel
/sad
It serves a purpose. Just because I dislike it doesn't mean I don't recognise that
Besides, it's still better than Supervised
It has a great purpose
To break
After automatically updating
So that the errors can be fixed for the rest of us
Some people run virtualization environments, and treat HomeAssistant like any other virtual appliance ๐คทโโ๏ธ
APC PowerChute and Cisco FindIT are similar. They both come down as a locked-down "appliance" style VM, provided directly by the OEM. Boot-and-run.
that's why it ends up in a container ๐ฎ
It handles its own containers just fine ๐คทโโ๏ธ
Thankfully I am not "some people" ๐
And if it ever borks, Veeam is backing up all virtual machines on the host, and can just roll the whole thing back in about 10 seconds.
which reminds me, I need to play with kubevirt at some point ๐ฎ
Containers have never worked well for me, especially when it comes to complex networking scenarios. ESXi also doesn't really do native containers (it spins up one VM per-host, and runs the containers inside that).
๐คท I've done some rather complex networked containers (in automation) before. It wasn't any more cumbersome than handling VMs
though k8s actually makes it easy
Caveat after discounting the heavy overhad and complexity that's just inherent to k8s
I've been through the ringer with Docker. I always wind up back on VMs in the end.
docker just kinda sucks
Networking (Docker's NAT, non-static addresses, etc.) always caused problems.
docker is pretty close to the "ship the devloper machine" meme
Unfortuntely, docker is what everyone seems to use these days
I'll often wind up spinning up a new VM to run a single docker container (looking at you Pterodactyl) because there's no good way to run it along-side other containers.
What does it do that breaks running it alongside others?
Granted, Pterodactyl takes over everything and spins up its own containers to run game servers, but still...
It's a management system for containerized game servers and game server clusters.
If you throw other containers on the same host, you're gonna have a bad time.
I suppose you could manually carve out resource allocation for your additional containers, but it expects dedicated "nodes" to run its own Docker containers on.
Think "kubernetes, but specifically for game servers"
apparently it takes control over docker.sock, that's like supervised ha
keep that to you and docker will be fine
It's easier to allocate resources if I just make a dedicated node VM for it to run its containers on, tbh.
in VM world that'd be the equivalent of giving a vm access to its hypervisor, of course its going to spin up shit
I did say it takes over everything and spins up its own containers ๐
Disgusting
It does a great job running game servers ๐คทโโ๏ธ
And it keeps each server in its own container, which helps with resource management and security
I've got it running Minecraft and Satisfactory servers right now
just run a dedicated docker instance for it then ๐ฎ you can do that both nested or on the host ๐คท
The current structure is vSphere cluster -> ESXi -> Rocky Linux VM > Pterodactyl
Can't run it on the host, as Pterodactyl requires docker, and isn't compatible with vSphere native containers.
As neat as it'd be to see the individual game servers it spawns as containers right in vSphere
did you know: you can run multiple docker daemons on the same host
its fugly, but it works
I've never been able to get that to work successfully, but more power to you
or just use containerd namespaces to emulate that ๐ฎ
I had a hard enough time getting Docker Compose to reliably set up my Plex/*arr stack
I run 70 or so containers, no issues here
Only VMs i run anymore are one windows vm to run Visio and one to run gns3
$ kubectl get -A pod | grep Running | wc -l
68
How'd you solve the dynamic IP issue? Docker loves to randomly give my containers different IP addresses, which causes some hell.
I don't, I use host networking
@dusky plank I converted your message into a file since it's above 15 lines :+1:
No
It's disgusting, but it works
Causes problems when Sonarr suddenly can't find nzbget because the IP changed
As an example
just use DNS o.0
DNS would be wrong because the IP changed
The DNS requires secure updates, and the containers are not domain joined
docker runs its own DNS
Don't confuse internal docker networking and DNS for external
Either way, never gotten it to work reliably. VMs always work ๐คทโโ๏ธ
what in the ๐ฆ is that?
i googled docker on steroids. was not disappointed
You found the new Podman logo. lol
Speaking of which, I've had a much better time with Podman, aside from some docker-specific tooling not working with it
for some reason windows people often get attracted by podman
I've only seen some strange austist in another discord that used podman, never seen it used in enterprise
just use systemd-nspawn at that point ๐
its heavily used by our linux devops, including artifactory
I swapped to buildah for building containers though
we have a slightly odd oneline around with kubectl run ... to start a container that's effectively a root shell on the host
nah
k5l, surely.
I had to look up a reference to A11y this morning
accessibility?
I mean, it kind of makes sense for i18n
what's weird about that?
there's two different spellings, after all
H8U H8U wouldn't D8U
lack of familiarity with that particular one and it's just a curious shortening strategy in general to me
yea, the shortening on those is kinda dumb as it's horrible for accessibility and discoverability
it's not just removing characters, but replacing them with a number that's completely non-obvious to the casual observer
This isn't HA related so I'll mention it here because y'all might think it's funny. I'm having an issue with my Enphase solar setup and it's probably got something to do with a Zigbee coordinator, since apparently that's what the Envoy gateway uses to communicate with the system controller.
heh
Probably because Windows and RedHat are both very common in SMB/Enterprise
And Podman is a RedHat thing
Think it was @late gate who had a lot of issues with his solar panels because they interferred with both his WiFi and ZigBee
Thankfully the only place I still see rhel is 5-10 year old power company appliances or oil companies that run legacy hardware
Ngl I was wondering if it might be caused by interference from my ZHA network when I figured out that's how those things communicate.
Yes, each panel talks to each other with ZigBee in some setups
I see RHEL everywhere man. I work at an MSP managing 90 contract clients, and most of the Linux VMs are RHEL.
It saves on support costs massively. Partially because each version gets 10 years of support, and partially because we can actually call RedHat for support if shit goes sideways.
And when you have one team of 15 engineers doing support for 90 companies, that counts for a lot.
Honestly, the weird part is how many of our clients jumpped ship from vSphere to Hyper-V
I'd consider vSphere the superior product, but I think cost is driving the decision more than functionality...
why bother? just run in cloud ๐คท
Because technical debt exists ๐คทโโ๏ธ
Some of these products and services can't be easily hoisted into the cloud. Some of these clients have HIPAA requirements to meet, etc.
Gotta love moving a service to the cloud, only to then need an on-prem server to back-up the cloud data because "one cloud backing up another cloud" doesn't cut it.
Yuck hyperv
And then you've got everyone who's been running Active directory on-prem for the last 20+ years. Transitioning away from that to Azure ADDS is... a process...
Thankfully I only see one customer every 5 years who run that
he, I need to ask our relevant team what the plan for backups is ๐ค
We went from 100% of our customers running vSphere/ESXi to only about 5 of them. Everyone swapped to Hyper-V the moment VMWare dropped support for their existing hardware.
"We can't upgrade to vSphere 7 without a new server? Screw that, I'll pay the $300 for a Windows Server 2022 Standard license key and run Hyper-V instead"
Hyper-V has hardware support going for it, I'll give it that much. It'll run on some very old servers with no complaint.
Only customers who run hyper-v are usually state, since they get it for free in their licence deals
It gets the job done, and keeps a ton of servers out of e-waste. I can't knock it too hard, even if it's not my favorite hypervisor ๐คทโโ๏ธ
I suppose the other reason they went the VMWare -> Hyper-V route is to preserve long-running backup chains. Most enterprise backup software supports both vSphere and Hyper-V, which means you don't have to swap backup products, and you don't nuke your RPO
The only other real choice that was compatible with existing backups was oVirt, and that's a tough sell unless you're doing a large deployment.
#ErCanEverything #DIY #PowerBank
๐Full Video๐ https://youtu.be/i73j_EMMoX4
To be fair, the only reason I get involved in hyperv cases these days is because they are crashing uncontrollably
About the worst issue we've had with Hyper-V is scheduling all the hosts in a cluster not to reboot for updates at exactly the same time. lol
Other than that, it gets the job done
Just off the top of my head
-vmqdepth causing storage issues
-cluster storage failover continuously
-rdma causing hosts to crash if io exceeds 100Mbit
-vm goes into sleep/suspend, causing physical core to do the same, causing host to bsod
-amd zen1 chiplet lulz causing a vm that exceedeing the core size of one chiplet to cause the host to bsod
I have noticed more issues on AMD-based servers than Intel-based ones, I'll give you that
Most of our clients are running Xeons. We have dual EPYC chips on our in-house server.
Had customers return multiple zen1 hosts and replaced them with intel hosts, all issues vanished
Getting the NUMA mode and threads-per-core balanced correctly was a pain
The whole chiplet nonsense makes coordination even more a pita
For some reason the server shipped in NPS4 mode, and nobody noticed until it was already in production
"Why is everything so slow? ...why are there 8 NUMA nodes???"
Fun fact, a single host running esxi, running a single sap vm
Default general purpose host Vs high performance hypervisor hosts power profile
18% increase in sap performance
With the latter power profile?
Yes
Brb, logging into iLO...
I'm pretty sure we're already set that way, but I don't trust anything anymore.
No need for that, just look at oneview ๐
Assuming it ever got registered with OneView...
Tsk tsk
Though apparently I just get redirected to GreenLake when trying to access OneView now
So that's fun
That just sounds like Greenlake for compute ops Management
Which is a totally seperate product from oneview
Well, blame HPE, because that's where I get sent when attempting to log in to OneView. lol
Then... Someone has done something funky in your management portfolio, we don't redirect you
Oneview can run fully offline and local
Honestly, this entire HPE account needs to be burned down and recreated. It's owned by our late CEO.

Oh, nevermind, you can actually change the account holder name. Just gonna set that to someone who's still alive...
But why. If something goes wrong, it's the perfect person to blame
Which reminds me, I need to figure out why he still shows up in Teams even though his account has been disabled for a while
"It's {{ states('sensor.random_name_of_coworker') }}'s fault!"
that should be a select probably
You want to pick the one to blame. A person with taste ๐
I dropped a 12v 9ah lead acid battery to the floor by mistake
Safe? It is not leaking
Don't lick it
Buttt
Anyone have a preferred brand of type A to C charging cables? (in USA)
Can't go wrong with Anker.
Its fine
Can an lxc be updated similar to an ordinary vm?
Would be a shame setting it up and then kinda being stuck
you mean HAOS?
No I mean a container in general
lxc is closer to a VM than docker o.0
But you can do updates in both container environments. in lxc you do them largely with a package manager in the container
with docker you'd usually keep all (Cache) data outside the container and swap it for a newer container for an update
You could do the same within docker, but it's just easier to rebuild it which is fairly easy
I kinda don't want to build an lxc template just to do that
you could. But docker is going to work against you with its concept of an (Container)image
im sorry I misread this entirely, yes you can update lxc... I read that as upgrade it TO a vm
still waking up
lxc containers usually contain an entire distro and run it similar to how you'd run a VM
while docker is more suited for application containers
lxc would be perfect for a dns server imho
I will see how it works out. Ordered me one of those beelink systems to both retire the last pi n possible my bigger proxmox host
nope
It better arrive soon hyped
Horstexplorer maybe this helps you understand a little better https://www.section.io/engineering-education/lxc-vs-docker-what-is-the-difference-and-why-docker-is-better/
there, i feel better. the Rpi heatsink/fan i ordered arrived. before install, the cpu temp was around 110F, now it is 85.5F and i think still falling ๐ not a huge difference, but i feel better. ๐
Ewww lxc
what's wrong with Rpi, it runs fine? it would be foolish to dedicate my 24-core rack server to HA
I do have respect for two ARM platforms though
UK met Isambard 2, and Proliant RL300
you wouldn't like a ARS-210ME-FNR?
Eh, my only real problem with them is the lack of paste in ipmi
I've thankfully only seen two supermicro systems in my life irl
And just the fact that they use a red led to indicate hot spare annoys me to smithereens
We tried moving to SuperMicros at work, it was going to be a huge savings compared to Dell.
Never managed to get our test cluster stable on them though. They had the weirdest and most obnoxious issues of any computer I think I've ever worked with.
They kept throwing MCEs, crashing, then on reboot the damn ipmi would freeze and we'd need to send someone to physically cycle them.
Lol
I'm convinced most people would be better with a cluster of used office Dells over a SuperMicro box.
They are hugely popular amongst homelab people who have no experience with proper hardware though
But mostly state side
Must suck for them.
Everyone I know with homelabs either has used Dell/HP/Lenovo servers or just repurposed desktop PCs.
I know one guy who loves them, he also argues with me that there are no 100% uptime systems
So
I've seen Dells booted for 6+ years, never seen a SuperMicro booted longer than a month.
lol when I quit my old job the server on my desk had 11 year uptime
I've downed servers with 13 years uptime, but I know of longer, but for an entire system, it can be upgraded fine online by booting individual nodes
The system is still online
Oh, yea, a cluster can easily have 100% uptime over years.
Modern storage systems are just clusters after all
it's like lego these days
I know they charge millions but I think IBM's top guarantee has like six 9's or something.
I've run 100% uptime phone systems too
for years
HA is not new
duplicated hosts, live backup recovery servers
You don't guarantee it like IBM does, though. With IBM their guarantees are more or less insurance policies, from what I've heard.
If the customer or end-user experiences less than 100% availability as a result of a qualified outage, HPE will work with the customer or end user to resolve the issue and provide a voucher that can be redeemed upon making a future investment in HPE Primera
That's not insurance
That is what the paper says, it happens differently
We even roll out new systems
If the customer system fails
Isn't that just warranty support?
Nope
Regular warranty stuff is just replacing parts until it works
A 100% uptime system would be replaced entirely in one go
We've had 6h CTR contracts for 20 years now
Where the entire system will be back up in less than 6 hours
I can't even imagine how much having a 6h replacement on standby would cost my place.
Fair
It matters quite a bit less if my customers have an outage, heh. It's not good but it's not compromise an economy bad.
Or oil and gas
Where a system outage gets expensive fast
When an oil engineer costs 2500$ per hour, plus lost income
docker aswell as lxc use the host kernel.
So they are both equally far from a vm imo. Canโt argue now have to board later aligator
yes
Hi all, I'm quite new to all of this, but I am using alarmo and node red on home assistant. I've created a scene that turns on all lights and sets colours ect when the alarm is triggered. Is it possible to create a variable or something, that when the alarm is disarmed/all clear, that lights will return to the state they were previously. i.e. if they were off they go back to off, if one room had lamps set to warm white they will go back to being on warm white etc. ?
Sure, save the state before the transition
Use that state in a scene
How often does your alarm go off? Best regards someone who hasn't had an alarm go off due to something else than user error the last 30+ years
Docker and LXC are user-space frontends to the same kernel-space mechanism.
They're just different configurations of the same technology
Both are just cgroups
Netham45 best man, confirms my flimsy facts
Lxc does manage it's libraries and applications less ephemeral afaik
Hi all I m quite new to all of this but
Yea, and Docker relies on overlayfs which is pretty closely associated, but those are filesystem layers, not virtualization/container layers.
Hopefully not often lol but I have the smoke alarms/carbon and door windows etc set to trigger a scene. How would I go about saving a state to a scene? And this is possible even with the state being potentially different each time? Thanks for the input. Learning as I go.
Hello everyone, what is the best way to remember current light (state, colour, brightness) and return it later in automation or script if required? I thought I would hold global variables as MQTT sensors, so in automation Iโm able to save state but Iโm unable to apply settings to the light when needed.
It was supposed to show 24h consumption, but it's now tracking the lifetime of my PSU in Watts because of my laziness to make it reset every 24h
now integrate it with a cost calculation and notice that... you still don't really care
At consumer scales you won't save much either way. Still it's pretty cool that I know I've spent 29.69โฌ for running my PC this year ๐ค
43c wasn't all that hot to begin with, tbh.
I'm dying ๐
I bet you're wishing that you had a water slide like #the-water-cooler message
Hahahaha
that's quite insane
Well without the 11.9mph wind rn we'd be dead
But it's 83% humidity
Sooo
Nasty hot
was a little chilly for my taste this morning with my windows open (~63F), but now it's a delightful 72F
San Diego has its own little microclimate
Just 29ยฐC, where is the problem?
At night, in an area where it doesn't get above 16ยฐc usually
Also that's the indoor temperature
๐
I'm getting cooked at +2ยฐc outside temp usually. So when its 31ยฐC like today, my room would reach 33ยฐC if i wouldn't have a heatpump.
The last few years, summer was really brutal, having peaks of up to 42ยฐC inside and no heatpump.
yeah I have no AC/heatpump either
im about to pull the circuit board from my AC to see why the relay isn't kcking the compressor on
do it now
I'm going with "pigeon crap"
Starter capacitor go out?
nah it's a split unit not a central
I used to keep spares when I had central AC
but it was also nearly $2K a month to run one of them, much less the both of them
On my Amazon Prime Day home page: Cat Exercise Wheel for Indoor Cats, Inside Diameter 39.37" 45.27โโH Large Cat Treadmill, Easy Assembled Cat Treadmill with Locking Mechanism, Ultra-Quiet Running Protect Spine,Reduce Vet Bills https://a.co/d/jcGcqwE
Product Description: Cat wheel, cat running wheel, cat exercise, cat wheel,, cat exercise wheel, help eliminate the lack of exercise for indoor cats Specification: Name: cat wheel Cat exercise size: XL (LรWรH): 11336120cm Runway diameter 100cm Cat wheel weight: about 24kg (Due to different meas...
๐บ๐ธ
Indoor cats, what a strange concept
34C at 0335
I almost died in my in-laws house last night when it was 22 in the daytime and no AC
LOL it won't be 22C here for months
We have an AC in the bedroom at home
Our master bedroom is in the sun wall and it gets to 30-35c inside in the summer
shit 22C is where I normally have my AC set at
Add four people to the mix and it is horrid without AC
4 people is a bit much for a master bedroom tho
Life with kids
I just did a small check of my data usage, I used 32GB on 4g last month (on my phone) and 80GB on WiFi last month (on my phone)
Think I used 40 or so GB more on the in car 4G probably
Home network was apparently only 1.51TB tx and 990GB tx
home internet only one way?
that would cost way too much here
Vnstat keeps track on both ways
Plus adds them
well it says tx and rx, you said tx and tx
Reminded me of my first mobile phone
Nokia 1610
Could only receive sms
Not send them
my motorola Ultra Classic knew nothing of SMS
That one is still top of the line in Bahamas?
lol clearly you know nothing of The Bahamas
folks here have the latest phones bro lol
I personally don't care to have the newest, I just 'upgraded' from my Galaxy S8+ to an S10+
but my kids got an iphone 14, or whatever the last one was
I wrote him out of the will for buying apple products
Ive always avoided company policy in favor of having something I know works
but now I make the policy
The policy is just they only pay for a new one every two years
I used to replace yearly with the old policy
my old company only gave $400 every 4 years for phone allowance
but somehow wanted you to change phones every 2
so I bought a $400 phone, and kept it lol
We usually replace laptop every three or four years, my last one lasted 5 though, but I used to replace every two years
and only $500 every 2 years for laptops
I told them I'm gonna buy chromebooks
and just throw them out when im done
lol
My current laptop was $3200 
my last new one was $2200
but that was 4 years ago almost
my current budget for a new laptop is -$1000
Yeah the old one was 2500, but that was in 2015
i like a good Raspberry Pi, they come in handy for some things. my standby generator monitor and HA "server" both run on a Rpi, just fine. but then, everybody is entitled to their own preference. ๐
๐คฎ
sorry but I ran on a pi for a long time and booooyyy should I have switched long ago
what advantage would i have running HA on a rack server?
rack server, as i have a unused Dell Poweredge R200 kicking around
is there much advantage of switching from my Rpi to my R200?
leaps and bounds difference going to even an old desktop
I'd personally avoid servers as old as the R200. It'll do more than your RPi but it's going to draw a ton of power to do it, it'll be noisy, and it won't really do much more than a more reasonable old desktop.
i have no income coming in, at the moment, so it is either the Rpi or the R200
An RPi 4 would give the 1950 a run for it's money performance-wise.
Those are easily replaced by any $100 desktop from AliExpress tbh
if i had $100...
If you don't have $100 then you don't have the money to power that R200, tbh.
i am not paying the power bills...
i am Autistic, and lost my job because of it. i live with my parents, currently
hmmm, my NAS has a Home Assistant plugin...
Plugin?
yeah, like a VM on TrueNAS-Core
it failed install though... gotta upgrade my TrueNAS OS
it would be cheaper for your parents to buy you a $100 desktop than to run an old server tbh
the TrueNAS server is already running anyway, thought it would kill two birds with one stone, but it failed to install the plugin
that's on an old PowerEdge R710 or R720 (can't remember model)
i doubt my parents would buy me a desktop... other than housing and utilities, i pay for my own projects (from odd jobs)
the odd jobs bairly pay my bills, so projects always take a back-seat
i will leave it on the Rpi for now, and maybe upgrade later
Truenas core is... Eh, jails were great fun a decade ago, but docker makes more sense
Scale is... Fine, if you hate to be in control
it works for my needs for now
with 18+TB of data, i am not moving to a different setup for a while
You can easily migrate a truenas core setup to any Linux setup in less than an hour, without migrating data or losing data
how?
Reinstall the OS drive with debian, do zpool import _poolname_ and your data is back
Then share data however you will
Hmmm...
zfs is very portable
good to know ๐
You should check out proxmox, too.
and ignore atx when he shits on it
Basically Debian + a management interface for running VMs.
You can do LXC and docker on proxmox no problem, too.
๐ค
just ignore atx in general
lol
hmmm, maybe when i can afford a newer desktop computer, i will redo my servers, and reduce to just the two. http/NAS/email and my FreePBX server
FreePBX runs great as a VM on Proxmox
The worst duty I've had at my current job was managing a FreePBX server.
Asterisk can go fuck itself.
runs fine for my needs. ๐
You're not running it across three offices across the country with a combination of copper pots, sip, and skype for inbound calls.
all i use it for is handling my "business" calls, and personal voicemail. all over old POTS
We used pots for local and sip for international
sounds like a reasonable plan.
lmfao
Asterisk is awesome
I used to love to use it to build custom conference bridges that have way more functionality than a $1,000,000 PBX
(and I have installed a few $1,000,000 PBX's)
I have FreePBX at home running my cisco ip phones
i have just two Polycom SIP phones, and a GrandStream cordless SIP phone ๐
shit that reminds me I have a cisco wireless phone someplace too
I mostly just never bothered to properly learn Asterisk's scripting language or flow so every time I had to do anything with it it was a fight.
ahhh now we are getting to the root of the issue.... sloth. LOL
lol
Eh, I knew that setup was on it's way out when I was asked to manage it so I didn't put much effort into it.
It's all outsourced now.
I ran FreePBX at home for a bit too so HA could call/text out when alarms/doorbells got pressed
it can do that? Hmmm...
I was using a plugin for it
now you can just call into the voice assistant which is nice
I'm gonna set that up ... eventually
gotta get internet at home back up and running first
When my doorbell rang it would call me and the people I live with and conference us to the camera above my doorbell
But the mic on that camera is pretty awful so it didn't work well
oh I have a doorphone
so that rings my hidden attendant when you press the button
so if I'm not home I can answer on my cell phone
Yea, mine would call me and say the doorbell has been rang, press 1 to conference in or 2 to ignore. When someone ignored it it'd hang up the call for everyone else.
And if they conferenced in I used that plugin to send the dtmf command to transfer the call to the conference line.
well my son never answers the door or phones
so there's never a need for a conference lol
infact he never answers when I call him either ๐คช
Nobody ever rings my door
yeah same
Just Amazon and UPS
same here lol
I wish, we don't get home delivery
I have my actual doorbell disconnected
Pressing the doorbell button triggers a zigbee relay that runs an automation to ping my phone
while its disconnected? that's ... scary
My dog freaks out way less now
Hmm?
I have the doorbell button wired to the switch line on a zbmini
And the zbmini is just powered with a 5v power supply
They say they take 120v but they'll run off of basically any power you give them
sorry I was reading too many things at once
I thought you had the button disconnected... I'm like what in the exorcist is going on at his house
Nah, the actual bell is disconnected
yeah that makes more sense
I had the zbmini triggering it but it was annoying having to calm my dogs down every time Amazon dropped off a package
I was thinking of making an alarm/notifier to put in my bedroom though. I was thinking a piezo on an esp8266 and I'd make it sound if the doorbell rang or any of my smoke alarms go off
Since I can't hear the smoke alarms on the far end of my house from my bedroom
so, if i install Proxmox on my PowerEdge R710 (24 core, 64GB ram), could i still import my ZFS pool?
or would i still be better to just get a newer desktop at some point?
Let me tell you about Norway, POTS has been officially dead since 2017, not sold since 2009, and now we are pulling copper out of the ground
Yes
well they are selling pots out the back of your modem here now
but you can get the SIP credentials from them if you wish
Yeah, we had pots via Linksys adapters on ftth in 2004
But I got rid of my pots line in 2011, because I hadn't had a call in 3 years, and the only one who called was my MIL
Or to be fair, my pots to IP adapter
Is what I got rid off
I haven't had physical pots since 2001
well the adapter is built into the CPE nowadays
I haven't had a POTS line or phone since I moved out of my parents house in 2010. It's still available as a service though.
Yeah, I migrated my pots number to sip in 2004
Android's built-in SIP client is pretty nice.
But never really used it
well being in telecom I've always had some form of voip
I still have some old H323 ATA's from linksys/net2phone but they took the service down ๐ฆ
Pap2T looks like the adapters we used to have
Think I threw out the one at my parents house in 2012 or something
Never seen that one
2 Cisco ATA-186's and a 188
and a SPA232 and SPA302 DECT phone
oh and I have an IAX ATA too
Thankfully POTS and other stuff died very fast
mostly pensioners who used it after 2005
I have a 2 port T1/E1 card for my asterisk PBX, just incase I wanna get funky too
well I only used them when I lived in Sweden
SDSL was the norm from 2000 going forward for most, if not already connected to fiber
everything is T-carrier on this side
for voice, not data
voice over T\E1 ?
really, never heard of that, probably because it was dead tech already when I started reading about it
lmfao it's not dead tech tho
carriers still use T-1 tie trunks a WHOLE lot
and it's pretty standard for voicemail systems to have them as backup
Y'all still use voicemail ?
define use.
if you mean request, pay for, and never bother to setup... then yes
if you mean have it, never check the messages then yes people 'use' voicemail here.
mobile services still include voicemail here, but I never check mine, and hardly anyone leaves me a message anyway, maybe 1-2 per year
my old provider sent me the voicemail as a mp3 file to email in 2009-2015, that was probably last time I checked one
I forwarded my unanswered mobile calls to a number that doesn't exist... so when the coverage activates it tells people my cell number doesn't exist
out of curiosity, anybody know how a PowerEdge T610 is on power consumption? i mean, if i can reduce all my servers down to just one...
I do like modern age with very little interaction, car service, online, new tires, online, order new internet provider, online, cancel any service, online
610 isn't great
but it is better than your 1950 and R200 combined
It's a full generation newer hardware
what about better than my R710?
610\710 is same gen
ok
so depends on what you have inside it
200\1950 is equivalent to G5 HP servers, which were released in 2006 istr
x10 is Equivalent to G6 \ G7 depending on CPU inside
so it would not be an upgrade from my R710
which are 2009 hardware
but also nehalem or newer
compared to the older nonsense CPUs
which are basically ewaste
maye i will reduce my 3 servers down to just my R710, to save power. and pay for it by selling the other servers on marketplace or something
*maybe
I ran on a 510 until last fall
it ran 4 VMs, 50TB of storage and 80 containers plus more fine
currently I run on a DL380 Gen9, equivalent to a R730, with 2VMs, 160TB storage and 80 containers and more
the server itself pulls about 180w average with 2x12c cpus and 128GB RAM
let me see if i can find what my R710 has for CPUs..
all storage and my lab server plus networking and router lands me at about 800w
says "Xeon(R) CPU X5670 @ 2.93GHz" X 24 cores in TrueNAS
Sounds about right
only 64GB ram though
I ran dual X5660s on my old 510
My Gen9 has dual E5-2620v3s
My lab server is a Gen8 with dual E5-2660v2s
Hmm, i mis-spoke before... my storage in it is 1.8TB, not 18TB... (slight difference) LOL
That is.... not a lot ๐
yeah, i know LOL
NAME SIZE ALLOC FREE CKPOINT EXPANDSZ FRAG CAP DEDUP HEALTH ALTROOT
Storage8TB 138T 42.4T 95.8T - - 2% 30% 1.00x ONLINE -
ohyeahmirrors 13.9T 269G 13.6T - - 0% 1% 1.00x ONLINE -
rustr1 2.17T 1.90T 282G - - 10% 87% 1.00x ONLINE -
i don't know why i thought 18TB
NAME SIZE ALLOC FREE CKPOINT EXPANDSZ FRAG CAP DEDUP HEALTH ALTROOT
Storage 36.3T 1.65T 34.7T - - 5% 4% 1.00x DEGRADED -
those drives i have are old though, i may want to see if i can back up the data and get newer drives...
Can get 10TB used enterprise SAS drives for โฌ100 each these days
I run a dRAID pool of 20x 8TB SAS drives for my primary storage
this only holds 6 drives total
maye i need, first, to save up and get a backup drive for the transfer...
that is easily expandable with a SAS shelf
1.8TB ? just store it in the cloud
that is so little data
i don't have that much cloud... more of a foggy fart really
Cloud is everywhere
i know
jottacloud unlimited is 5TB before capping, costs about $8 per month
same for google drive, 2TB is about the same cost
backblaze is even cheaper
my google drive has 10GB... not playing with huge funds here ๐
Never seen anything this. Simply wow
Since google drive comes default with 15GB, that is impressive ๐
You've never seen zfs before @tidal bronze ?
oh, your correct, it is 15G LMAO
Things not to do: Run dedup with big files on a zfs nas, on spinning rust.
not without a special vdev, at least
Never seen, how frustration of isolation and staying inside 4 walls, 10 months of the year in TB. Interesting and cool.
dedup is fine, with special vdev yes
I just do compression though
might do dedup with my new SSD pool
"delete this 20TB of files please". "Ok, but I'm going to lock up and stop anything doing anything when I'm doing any file over a terabyte."
pulling in 32 2TB SSDs
running on 16TB spinning disks.
Yeah, my primary pool is rust
fast enough for anything I need it for
but I need a scratch place for playing around
I shouldn't have been deduping or compressing anyway. it was all backup so already compressed. And the likelyhood of dedupe was low. especially when some was encrypted
(dedupe kicks in after encryption on zfs)
there, between my laptop drive and my foggy fart (google cloud storage), i am getting backed up...
cleaning out my NAS server, so i can re-do it
kinda exciting, re-doing a server ๐
I don't think I've redone my backup server since 2009
but my primary system has been migrated from linux to BSD to linux at two points between 2007-2021
1.8 footprint only?
yes
Running lean I would say ๐
yeah, big drives were out of my price range at the time
Oh I was not referring to storage just actual data. Storage size means very little. ๐
storage size is 2TB
data is 1.8TB
mostly just raw gopro videos for my YouTube channel...
Storing other people's content in TBs of data is kinda pointless. So you are running a lean operation ๐
just my own data, other than my website with forum
hopefully after all this, i can figure out Proxmox so i can get back up and running...
to be fair untrained chimps could prolly figure it out
๐คฃ
it's nothing like Xenserver for example
Xenserver?
yeah
i know of XenForo, but never heard of Xenserver...
it's a hypervisor
I had a Xenserver on my desk at my old job with an 11 year uptime
pretty much the same, except trade autism for dyslexia and ADD
I do have a collection of certs tho but mostly self taught myself that stuff too
only cert i have is the SSL cert for my webserver ๐คฃ
oh boy
and i botched that a little
I am probably a high functioning austist to be fair
or some other spectrum diagnosis
I can learn mostly anything in an hour or two
yeah that's an ADD/ADHD trait also
and I prefer working with multiple things at one time
the more customers I work with at once
the better I work
i don't do with customers, i don't like people ๐
yeah same, I had triple monitors and 2 laptops
in IT, working with customers doesn't mean you have to see or speak to anyone
it's the general term we use when logging into machines
They are beautiful!
I also do not like humans.
i know, i was tech support for Prudential Securities for a bit
funny thing is, i am a carpenter by trade. computers is just a hobby
I hope cats take over the world, and keep me as a collaborator
I'm a tech geek for fun, for work, I just get paid to do my hobby
but I do carpentry for fun too
same, me and another IT buddy just built a 16ftx18ft shed
pick up a hammer and drive a nail. if it sinks to the head, your a carpenter ๐
ewwww hammers
Lol. I am involved in loads of remodeling etc but I don't get to do it.
like, use a screw gun and screws, what you think this is... The Oregon trail using nails?
im also a fan of joinery without having to use nails or screws
๐
what about remote jobs where there is no power yet, or power fails?
I live in The Bahamas, I expect there to be no power... I take a generator and cordless tools everywhere
I even have a 1500W inverter in my car
I doubt many know how to even hold a screwdriver anymore. ๐
<-- does
I still have a hand crank drill that was my grandpa's
so I don't need power tools even ๐
Power tools a must now, to save time but sad the hand skills are dying. Oh well, life!
well I keep them alive
I also still have grandpas old coping saws and chisels
but sadly don't have much time to play carpenter cuz I need to do all the shit I get paid to do first lol
Haven't seen a chisel in ages.
I have half a toolbox full
Ever in doubt use an impact gun I guess.
The new tools they come up with is mind numbing to say the least.
iwish the batteries had a longer life. i have many $100s in batteries that will never take a charge again
rebuild them
Hilti or Milwaukee, end of story.
Makita, they have a chip in them
๐
Hmmm...
But why?
Time is not
I just do stuff on company time and get paid to do other stuff ๐
I even use the bathroom on company time so I don't have to do it on my time
That's sad
company pays me anyway, if I do other stuff when I have free time, perfect
then charge overtime when I have to work late
๐
Common with people who hate their job
I love my job
Or bad at what they do
my job is my hobby
I haven't worked for anyone in 25 years I wouldn't know how to steal time. But sad to hear people dong that.
I don't steal time LOL
I'm self employed also
but somethings you can't exactly quantify in IT... when I get an idea at 3am, who pays for it?

You gotta be able to walk the talk. Easy to lean on some blind US company to pay your dues but hard to actually do it.
<-- self-unemployed ๐
Brilliant
self-employed sometimes means self-unemployed
Hard truth
works been slow AF here, now everyone seems to have remembered my number
and I can't even cut my grass in 4 weeks
Good for you!
There you go!
My calendar is mostly full these days, which is good, I'm tired of doing certifications and trainings
18 certs last year
5 so far this year
My company's attitude seems to be they don't care how many hours I work as long as stuff gets done. It's nice.
that is the way it should be
that used to be how it was at my old job
I would get a days work done before 8am and just do what I wanted the rest of the day
then they hired managers
Epic
EU managers understand how hours vs output works, US appears to not do
I don't believe in time management. But if I had free time and someone else was paying for it, ๐ฏ I am not doing something I love to do.
Yea, I've woken up at 3 AM, did all my work before sunrise and worked on personal projects for the rest of the day before.
cept I'd wake up at 6am, and work on 6-8 customers at the same time and be done by 8am with a full days time billed
Yeah, I've easily billed 35h of work in a single day
How do you guys have free time? Is that a thing?
still have to wait for those processes to finish
I was working 20 hours a week and billing 80 LOL
I have lots of free time
mostly after 8pm for me
actual hours worked is probably 20 per week realistic for me
Ive been working 7 days a week for 4 weeks tho
the rest I spend with the kids and around the house
Doing that 20+ years
well truthfully so have I, but normally I have some time to do shit... I've not had time at all recently
getting closer to my plan of moving rural and work 100% remote together with the wife now
she is also looking for 100% remote with her manager
get so much more done
and since most of my current projects are international it doesn't matter where I am located
I think having to go to an office to work in IT, is pretty stupid
the 2 hours I'd sit in traffic I could do all the work I need to for the day
exactly
losing 4h of work every day
at least for me, I don't get up until 8am anyway 
I think I'm going to start calling our datacenter guys disk jockeys.
1985*
yeah, I haven't seen a non all flash system since probably 2015
We still use spinners for our backups, that's it
Investments! ๐ you will have more time do more investing. Goal would be to make money while you are sleeping ๐ด
Hmmm, when i re-do my servers tomorrow... somewhere i gotta find a monitor keyboard and mouse to set it up... ๐ค forgot about that little detail...
never used it...
time to
Any form of LOM is damned handy
At least, if it doesn't use decade old Java crap
oh no we have awoken the Tinkerer
I hate double meaning abbreviations
I know all the script kiddies just say IPMI, I prefer to say OOBM
Who really thinks lan on motherboard when they hear lom?
atx does
FLEXLOM = Flexible Lan On Motherboard
OOBM can be serial too ๐
LOM in HPUX is the onboard cards
Is the flexible one on a bendy PCB or something?
ahh the days of NIUX
iLO is serial too @steel crag
yeah but I mean mostly on cisco stuff
do i need special software to connect to iDRAC?
special onboard slot that connects directly to the systemboard instead of riser
Nopes
iDracs can connect to a virtual serial port on the host if you ssh into the drac
x10 iDRACs is unfortunatly java bullshit
but there is an iDRAC container
just run that container on your desktop or whatever
There's been html5 interfaces back to the 12th gen iDracs
Or just use a server made after 2009
Me like iLO over iDRAC
iLO is superior
Lenovo's TSM sucks, SuperMicro's BMC sucks more
especially with the restful api interface, that lets you create RAID configuration etc before even powering it on
and it will apply it after powering it on
great for remote deployment
same with firmware
You can queue actions like that on dracs too
Yea, I used it to schedule some bios setting changes last week
It was on an iDrac 9
yeah, we could schedule bios changes from iLO3, so 2009
Job queues is what they call it
but stuff like RAID config and changes to iSCSI config etc was 2015 I believe
Idk when it was added, I've only looked on newer boxes.
I used the bios stuff to deploy BIOS settings to 800 HPC nodes in 2011
was pretty sweet
making all nodes equal in two commands
all nodes are created equal, just some nodes are more equal than others
yaaas queen
Orwell is rolling in his grave
๐คฆโโ๏ธ
is there some special sauce to mounting an nfs share into an lxc?
proxmox needs a subscription $$$?
would have been nice to know before installing it...
only if you want their enterprise repo stuff
lol
you just disable the enterprise repos and add the pve-no-subscription one
there is a script to disable the disclaimer aswell if thats what bothers you
ok...
does anyone know about a 1U rack with space for multiple 2.5 U2 drives?
Pretty sure most HPE servers from Gen9 forward have U2 or U3 support
how many slots depends on backplane
why mount it inside the container @solar zealot ? why not mount it to the host OS then map it to the container ?
I don't touch LXC though
i supposed this would be the right way as its not something the host would need access to in its entirety
ok... it is not letting me import my ZFS pool
i am on the web interface, there was no iDRAC installed
web interface of what ?
proxmox. Click your server, click shell.
proxmox
it seems like i would have needed a privileged container to begin with
works :3
then zpool import <poolname>
ok, i had to do zpool import -f Storage
probably because it wasn't exported on the original host
sure, scp it in
or use the upload in the web ui
yuck webuis
do you use arch btw 


๐