Werner changed the topic of #armbian to: armbian - Linux for ARM development boards | www.armbian.com | Github: github.com/armbian | Commits: #armbian-commits | Forums Feed: #armbian-rss | Type 'help' for help | This channel is logged -> irc.armbian.com
<lanefu> awf_sbrt: definitely possible unfortunately
<awf_sbrt> Is there any good way to fix it that won't cause issues in the future? I could try to make the initrd smaller (I'm not sure what caused it to grow so much), or I could change the load addresses, but I'm not sure which option is better.
fizikz has joined #armbian
<fizikz> is this normal? in omv5 i see a pending update for armbian-firmware 20.05.5, but the armbian changelog shows 20.05.5 as never released/skipped: https://docs.armbian.com/Release_Changelog/
ChriChri_ has joined #armbian
ChriChri has quit [Ping timeout: 260 seconds]
ChriChri_ is now known as ChriChri
<lanefu> awf_sbrt: do you have a dedicated boot partition
<lanefu> fizikz: maybe that's cached. Try an apt clean and apt update
<fizikz> lanefu: at this point i already installed it successfully. confused about the version number vs the changelog notes though
<fizikz> and someone else confirmed the same 20.05.5 version on their device, so it's not just me
<awf_sbrt> lanefu: No, and I'm starting to figure it out. I managed to reduce initrd to <13MB so it should be able to boot now.
<lanefu> awf_sbrt: how'd you clean it up
<awf_sbrt> Removing nbd initramfs hook saved a little over 2MB. It's still a bit bigger than I'd like, so I'm trying to fix that.
<awf_sbrt> (I was playing with nbd at some point. I don't need it on boot.)
<awf_sbrt> It still pulls in more stuff than I'd like, so I'm looking at that.
<lanefu> fizikz: honestly our release methodology is very much a WIP so its possible .6 wasnt built for all flavors
<fizikz> lanefu: but 5 was? even though the changelog says it wasn't released?
<lanefu> what device
<fizikz> odroid hc2
<lanefu> okay so yes armbian firmware package is version as 5 https://apt.armbian.com/pool/main/a/armbian-firmware/
<lanefu> the kernel should show as .6
<lanefu> sorry for the confusion around the package versions
<lanefu> tldr 20.05.05 is correct for armbian-firmware
<awf_sbrt> Okay. I'm around 10MB now, slightly smaller than on a fresh image. Rebooting now...
<lanefu> awf_sbrt: sweet. so the hooks are litraly upset because the package is too big, right?
<fizikz> lanefu: so armbian and armbian-firmware have different versions?
<lanefu> yes
<fizikz> kernel should show .6? mine is 5.4.41-odroidxu4
<awf_sbrt> lanefu: U-boot loads the initrd at 0x43300000, and after that it loads the fixup script (and overlays, if I had any enabled) at 0x0x44000000. That leaves 13MB for initrd, if it's any larger, it will overwrite a part of it and cause a checksum error.
<awf_sbrt> Disabling lvm2 and mdadm hooks saved another 3MB, which brought me close to a fresh image which don't have those.
sassinak-work has quit [Ping timeout: 240 seconds]
sassinak-work has joined #armbian
<lanefu> awf_sbrt: ahh okay! that makes sense..albeit unfortunate
<awf_sbrt> lanefu: Mmhm. At some point the load adresses will have to be changed I guess, it's only going to be a bigger issue as things grow. But now my system is working, and I have working 3D acceleration!
<awf_sbrt> And thanks for your time!
<lanefu> awf_sbrt: actually the kernel itself probably won't show that.... the versioning is primarily oriented around the published images
<lanefu> awf_sbrt: awesome you got 3d working... i need to do more on my desktop boards... anyway you find a really important edge case I'll try to log it somewhere
<awf_sbrt> I don't think the first message was meant for me!
<lanefu> s/actually/fizikz: actually/
<ArmbianHelper> lanefu meant to say: awf_sbrt: fizikz: actually the kernel itself probably won't show that.... the versioning is primarily oriented around the published images
<awf_sbrt> Okay! I'm glad I helped with something.
<awf_sbrt> Oh, I also disabled the busybox hook, since it's not on fresh images either.
<lanefu> fizikz: the xu4 image is probably still 20.05.4...
<lanefu> awf_sbrt: yeah there's a lot of workto be done on the upgrade front :(
<awf_sbrt> lanefu: Mmhm. This is started as a Jessie install if I remember correctly, so I've hit a few issues over the years. But I always managed to fix it.
<lanefu> awf_sbrt: yeahhhh you've had a good run
<fizikz> lanefu: the xu4 image is probably still 20.05.4 but armbian-firmware 20.05.5 package was installed? getting really confusing...
<lanefu> yes
<lanefu> I don't know how to put you at ease, but don't worry about it. as mentioned earlier the release process is still a WIP
<fizikz> not worried, since i haven't experience problems due to this. just confused
<lanefu> cool, yeah sorry for the confusion
<fizikz> on another note, any thoughts on why my weekly "borg check" went from taking ~2.5 hours on omv4/armbian kernel 4.y to 5+ hours on omv5/armbian kernel 5.y ? notice the load spikes are higher on omv5/ kernel 5.y (data since may) https://i.imgur.com/AYiMU52.png (note borg is single-threaded)
DaRock has joined #armbian
<lanefu> fizikz: eww that's a bummer... any major changes in the borg version?
<lanefu> whats ` cpufreq-info |fgrep decide|tail -n1` say
<fizikz> the borg version is the same 1.1.9
<fizikz> The governor "ondemand" may decide which speed to use
<fizikz> lanefu: in case it helps, here is the armbian-config benchmark with omv5/armbian kernel5.y that i did a while back: http://ix.io/2k57
<fizikz> i just noticed something right now: is it strange that section "System health while running 7-zip single core benchmark" is showing cpu loads of up to 7?
<lanefu> 7z is going to run threads based on number of cores.. and since xu4 has 8 cores tha tmakes sense
<lanefu> the whole big-little thing is kind of voodooo.. maybe make sure borg's pid is bound to a big core?
<lanefu> oh while running single core.. at 7
sunshavi has quit [Read error: Connection reset by peer]
Tonymac32 has joined #armbian
<Tonymac32> woof
sunshavi has joined #armbian
andy1983 has quit [Quit: Leaving.]
<lanefu> hey Tonymac32 !
<lanefu> fizikz: you've got me curious.. so running benchmark on a mc1-solo
sunshavi has quit [Read error: Connection reset by peer]
sunshavi has joined #armbian
<fizikz> lanefu: thanks. solving this would be a load of my mind
Dr{Who} has quit [Remote host closed the connection]
Dr{Who} has joined #armbian
<Tonymac32> I had a faulty liquid cooler on this machine of mine, found it at full stop a few weeks ago, when rebooted, got 800 MHz 1 core
<lanefu> smoooookin
<Tonymac32> so I bought a massive heat pipe heat sink with a 140 mm fan
<Tonymac32> it's now almost impossible to break 45 C
<lanefu> haha
<Tonymac32> I am really disappointed, it was a Corsair H100i closed loop setup, my H60 is still plugging away 7 years or so after I put it on my overclocked 3770
<Tonymac32> but, since I run full towers, there's really no reason to spend that money I see now.
<Tonymac32> So now 64 GB RAM and a cooler so big I was surprised the case shut on it, all win
<Tonymac32> ok, to topic: Meson64 5.7 has no audio again, hahaha
<lanefu> Tonymac32: meson can kiss my ass.. im still fighting 4k even wit 5.7
<Tonymac32> interesting
<lanefu> it detects my monitor higher resolution modelines..but for 0hz refresh
<Tonymac32> did you build with no patches? I think we are still fucking it up somewhere in our stuff, since it's a convergence of a lot of other work
<lanefu> Oh thats a good idea
<Tonymac32> when you take libreELEC and Khadas and upstream patches that haven't been accepted all at the same time, it gets ugly really fast
<Tonymac32> that said, I have also found that my cables can influence whether my 4k works or not, even "good" cables seem dodgy sometimes
<lanefu> yeah so im in a dumb situation where im using an ACTIVE hdmi 2.0 to display port 1.2 converter
<lanefu> ...4k works fine with kernel 4.9
<Tonymac32> oh my
<lanefu> but the allwinner stuff handles it no prob
<lanefu> i even took my display port kvm out of the equation
<lanefu> at least 1080p is working.. i was stuck at 1920x1200 which isnt the best aspect ratio for upscaling
<Tonymac32> hmmm
sassinak-work has quit [Ping timeout: 260 seconds]
sassinak-work has joined #armbian
<Tonymac32> "HACK", "WIP", "FROMGIT", "FROMLIST" All in one pile in this patch directory. :'(
<Tonymac32> *puffs on cigar, puts on boots*
<lanefu> hahahaha
<lanefu> maybe we should complicate things more by putting patches in a database and then dumping them by queyr
<lanefu> fizikz: mine did same thing on single core benchmark
<fizikz> lanefu: interesting. that doesn't seem right, does it?
<fizikz> did anything else stand out?
<lanefu> wasnt watching close enough
<lanefu> lemme just run that one benchmark by hand
<fizikz> you didn't save the benchmark link?
<Tonymac32> wait, what's it doing?
<fizikz> single threaded benchmark has loads as high as 7
<lanefu> fizikz: yeah i stupidly ran it through armbian-config and didnt get a link
<fizikz> lanefu: it's been a while since i ran it, but i thought i used armbian-config and got a link printed in the terminal. unless i don't remember correctly..
<lanefu> fizikz: okay the process still threads, but it only crunches on 1 core https://snipboard.io/6gDXvV.jpg
<Tonymac32> I need to retest running high loads on those boards, it would hang on kernel 5.4
<fizikz> lanefu: could that indicate a problem or is that just an issue with the benchmark?
<fizikz> interesting that htop shows you the big and little cpus with different frequencies. on the hc2 all 8 are incorrectly shown as LITTLE
<fizikz> Tonymac32: i've been having an issue on kernel 5.4 /omv5 where "borg check" takes ~2X as long to run than it did on kernel 4 / omv4. not sure why, but this strange benchmark result caught my eye
<lanefu> fizikz: try wrapping your borg command with taskset -c 7
<lanefu> that should force it on a big core
<Tonymac32> I don't know how it would misidentify on the HC2 vs the HC1, they are literally identical
<Tonymac32> They should even share the device tree
<fizikz> lanefu: i did watch the borg process when it was running and it stayed on the big cpus
<fizikz> my comment was about htop not identifying cores/frequencies correctly
<Tonymac32> which is *really* weird, for reasons mentioned above
<lanefu> fizikz: what version of htop? mines 2.2.0
<fizikz> yes, 2.2.0
<Tonymac32> what are the 2 kernel versions/etc
<Tonymac32> I've never seen what you're describing
<Tonymac32> (XU4 and MC1)
<fizikz> i have only the HC2. it's on kernel 5.4.41-odroidxu4
<lanefu> fizikz: having same behavior on single core test with my odroidn2.. so anyway red herring
<lanefu> HC2 is just an mc1 with a sata chip i believe
<Tonymac32> Right, the MC and HC series share boards
<Tonymac32> the MC is a stripped down HC
<Tonymac32> Well, the HC2 is for the bigger boards, it might have a different power config
<lanefu> fizikz: whats borg doing... have you been watching file IO... `iostat -mx 5`
<lanefu> io scheduler might be different between the kernels
<fizikz> it's not running now. but if you suggest things to check i'll try them out next week when it runs
<lanefu> whats the actual task doing? hashing? crypto?
<fizikz> i'm not too well-versed in this, but here are some details: https://borgbackup.readthedocs.io/en/stable/usage/check.html
<fizikz> someone else is running the benchmark on the C2 (4 core) and it finished much more quickly than the HC2 (8 core). is that expected?
<Tonymac32> no
<Tonymac32> well, let me * that
<Tonymac32> there are some accelerations available to the 64-bit S905 that the 32-bit exynos does not have
<Tonymac32> but
<Tonymac32> the A15 is a MUCH more powerful core than the typical A53, and is clocked a lot higher
<Tonymac32> sudo apt upgrade
<Tonymac32> wrong window
<Tonymac32> ROFL
<fizikz> i also noticed that the memory benchmarks are hugely different
<fizikz> latency and bandwidth are multiple times worse on my HC2 than that C2
<Tonymac32> that one is a bit tougher, and with Amlogic magic blobs I have no idea what the RAM on a C2 is doing
<fizikz> my results: http://ix.io/2k57
<Tonymac32> I'm on 5.4.19 at the moment
<Tonymac32> hang on
<Tonymac32> lanefu that memory bandwidth is bad
<fizikz> lanefu: wonderful. thank you. so far i've been begging other HC2 users to bench theirs for comparison. one is currently running
<Tonymac32> memcpy of 85 MB/s
<fizikz> yes that shocked me for memory, but i had no comparison
<Tonymac32> you should be getting ODROID XU4 numbers
<Tonymac32> so 2200-ish
<fizikz> also note the latencies. they seem really high to me even without a comparison
<Tonymac32> I'm going to test my server
<fizikz> Tonymac32: it does get 2400MB/s for memcpy on the second run down the results page though
<Tonymac32> yeah, but the little cores shouldn't see that big of a hit
<Tonymac32> you can access a good SD card that fast. ;-)
<lanefu> okay kicking off fresh sbc-bench run aagain
<fizikz> yeah. even hdds are faster now :S
<fizikz> i got 200 MB/s + on a 10TB HDD
<lanefu> Tonymac32: so yeah memcopy onthis kernel 4.14 test from tk's archive def faster than 85 http://ix.io/1iLy
<Tonymac32> 4x faster
<lanefu> yeah
<fizikz> hmm
<lanefu> Tonymac32: actually if you sitll have 5.4.19 run sbc-bench on that
<fizikz> i might retest just to rule out weird inconsistency
<Tonymac32> that was my goal, but appatenly my server is thinking about something too hard
<Tonymac32> need to kill some processes
<fizikz> also, do these boards have reputations for dying? or other weird problems?
<lanefu> fizikz: teh default response to weirdness is power quality, and sd card quality
<lanefu> for any SBC weirdness
<Tonymac32> not that I'm aware of, I've been using an XU4 as my media server for 5 years now
<Tonymac32> well, I found my failed attempt at setting up nextcloud was hogging some processor time, that's fun
fizikz has quit [Read error: Connection reset by peer]
fizikz has joined #armbian
<fizikz> irc client (pidgin) crashed :S
drobo_00 has quit [Quit: drobo_00]
<fizikz> micro sd card is a 64gb samsung, and the power supply is a 12V/2A from ameridroid
<Tonymac32> how much does the drive need
<Tonymac32> for reference the XU4 comes with a 5V 4 Amp supply
<Tonymac32> 20W with no drive
<Tonymac32> you have 24W with a drive, and probably higher losses due to regulators
<Tonymac32> I won't hang my hat on your power supply just yet though, let me test my XU4 which has a 10A chassis supply :)
<fizikz> hmm according to the product brief the WD100EFAX can take up to 5.7W on read write
<fizikz> 1.79A peak
<fizikz> another hc2 user got 84MB/s on small core memcpy too: http://ix.io/2qo4
<Tonymac32> yeah, I think this is a kernel problem, but you might want to get a bigger power supply anyway. ;-)
<fizikz> i hope ameridroid knew what they were doing when they sold this powersupply for the hc2!
<Tonymac32> ... As an EE, I will propose to you that they did not
<lanefu> lol
<fizikz> lol ouch
<Tonymac32> the 4x A12 cores at full load can probably pull 10-15 watts
<Tonymac32> so add 4 small cores, the rest of the board, and the drive
<Tonymac32> you never want to be at 100% of your supplies capability
<fizikz> i have a 12V/4A power supply for an external hdd dock that i could try, but i'm not sure if the connector/polarity is the same
<Tonymac32> center positive is almost universal, other than some withcraft voodoo worshipping by guitar people
<fizikz> but also, i noticed the performance difference between omv4/kernel 4.x vs omv5/kernel 5.x so ... maybe more likely the issue lies there?
<Tonymac32> like I said, I think we have a kernel problem, but your supply will probably let you down eventually
<fizikz> good to know. or bad to know. or ..
<Tonymac32> bad to find out, good to know before bad things happen
<fizikz> yes
* lanefu still waiting on benchmarks
<lanefu> on 7zip now
<Tonymac32> think about it this way, that external hard drive has the same type of disk in it, and nowhere near as powerful a controller
* Tonymac32 also benchmarking
<fizikz> only detail is that external hdd dock is a dual slot
<fizikz> so 2 X HDDs
<fizikz> potentially
<Tonymac32> right, and it still doesn't have a linux computer in it
<Tonymac32> probably a USB hub to 2x USB-SATA, right?
<fizikz> yes
<lanefu> Tonymac32: would be impressed with my 8 SBCs running on a USB charging station
<fizikz> it has a single USB cable. not sure how it is built internally
<Werner> Good morning
<lanefu> morning
<Tonymac32> Guten Morgen
<Tonymac32> so my memcpy is shit too
<Tonymac32> 80.2 MB/s
<fizikz> woohoo, you're in the club
<fizikz> another hc2 user also confirmed similar results
<Tonymac32> I can hear tkaiser throwing things in Germany
<Tonymac32> ok, that was 5.4.19, apt-upgrading
<lanefu> haha
<Tonymac32> of note: my HTOP doesn't specify what cores are what
<lanefu> yeah it takes a fresh htop and the right voodoo
<fizikz> no? what does htop show you?
<Tonymac32> no frequencies
<Tonymac32> but it hasn't been updated in a long time, and is BIONIC
<fizikz> the other omv5/hc2 user has htop showing the cores correctly (4 little, 4 big)
<lanefu> Memory performance (big.LITTLE cores measured individually):
<lanefu> memcpy: 78.3 MB/s
<lanefu> memset: 278.4 MB/s (0.2%)
<lanefu> memcpy: 2260.8 MB/s
<lanefu> memset: 4917.7 MB/s (1.0%)
<fizikz> it's one big miserable family :)
<Tonymac32> hahaha worse than me
<Tonymac32> :P
<lanefu> alright im gonna celebrate Werner's awaking by going to bed
<lanefu> later
<Werner> Sleep well
<Tonymac32> haha later
<fizikz> good night. thanks for the benchmarks
fromport has joined #armbian
<Werner> Tony, do you want a project affiliation cloak? Like @armbian/maintainer/Tonymac32
<Werner> The 2nd term can be chosen more or less freely, like contributor, developer, staff...
<Tonymac32> That's up to management. ;-)
<Werner> Also feel free to take a look at the irc section of our documentation to get a overview if you haven't already ;) .rtfm
<ArmbianHelper> Check out our awesome documentation! It's tremendous, promise! https://docs.armbian.com/
<Tonymac32> rtfm rofl
<Werner> Yes and no. I have to ask users if they like to have a cloak. I am not allowed to do this on my own.
<Tonymac32> sure, it would be cool if the cloak came with invisibility or something, but affiliation will do. :)
<Werner> What do you mean by invisibility?
<Tonymac32> "cloak of invisibility", like magic.
<Tonymac32> it was a joke, nothing serious
<Werner> Ah haha. Yeah well not that kind of cloak ;)
<Werner> Just let me know what you think which term fits best
<Tonymac32> maintainer is fine.
<Werner> Okay, will do that.
<Werner> Have to communicate to freenode staff and depends on when somebody has time ;)
Tonymac32 has quit [Changing host]
Tonymac32 has joined #armbian
<Werner> Done
<fizikz> Tonymac32: found it. hardkernel recommends 12V/2A power adapter and says the hc2 consumes less than 1A/12V in most cases, and up to 2A with high load and spinning 3.5" HDD
<fizikz> see section "which power adapter should i use?"
* Tonymac32 has skeptical face on
<fizikz> but as you pointed out, if it draws up to 2A and the supply is 2A...
<Tonymac32> yeah, I think we have some threads talking about power supply aging/etc
<Werner> Running stuff on their limits is not necessarily a good idea.
<Werner> Learned that recently at work the hard way :D Had a relay that was rated for 6 amps and had a constant flow of 5.2 amps. Well, end of story was that it melted some day ^^
<fizikz> makes sense. but given the benchmark results now, from at least 4 devices, i think it's reasonably clear it's not a hardware issue
<fizikz> melted :S
<Tonymac32> Chrysler designed their fuse boxes with a relay on a PCB to run the fuel pump, because the fuel pump relays that plug into the sockets never fail
<Tonymac32> well
<Tonymac32> the surface mount relay meant for telecom use does break
<Tonymac32> because it's rating is not high enough above the motor inrush current
<Tonymac32> and it can't be serviced
<Werner> Broken by design...like the Allwinner H6 ^^
<Tonymac32> haha they list the TDP of the exynos as 5W. using that I can see where someone would imagine a 2A supply adequate
<Tonymac32> Werner :'(
<Werner> For the H6 it needed a smart Chinese woman to get around it....not sure what to do in Tony's case :/
<Tonymac32> hahaha
<Tonymac32> I literally bypassed it with another relay
<Werner> xD
<Tonymac32> freenode said stuff to me about hidden hosts, so they had time I guess
<Werner> Yep. It has been applied. Simply do a whois on yourself.
<Tonymac32> so I wonder if this complete fail of memory bandwidth had something to do with my BOINC task crashes on the MC1
<Tonymac32> I noticed my server wouldn't transcode anymore but didn't think much of it since it was a shaky software deal anyway
<fizikz> Tonymac32: who listed exynos TDP at 5W ?
<Tonymac32> it popped up in a couple 2nd hand data listings. I would imagine this is linked to cell phone usage
<Tonymac32> hahahahaha I know
<fizikz> well i guess today's "accomplishment" was making others aware of their devices' miserable performance. for a while it seemed like i was the only one who noticed.
<Tonymac32> I need to learn more about HMP apparently, I did not expect there to be a real RAM access difference outside of clock speed
<fizikz> do you think these results would explain the increased time for borg check?
<fizikz> the little core memory speed difference is significant between kernel 4.x and 5.x, but the big core seems similar
<fizikz> i redid the benchmark in case something changed in the past 2 months: http://ix.io/2qom
<fizikz> same thing, a bit worse
<fizikz> memcpy: 79.3 MB/s; memset: 278.5 MB/s (0.1%); memcpy: 2315.5 MB/s (1.0%); memset: 4884.3 MB/s (0.8%)
<fizikz> 7-zip total scores (3 consecutive runs): 7286,7152,7217
sunshavi has quit [Remote host closed the connection]
<Werner> sunxi64 on 5.7 broke cpufreq :/
sunshavi has joined #armbian
<IgorPec> good morniung
<IgorPec> werner: well, it would be amazing if everything will work OOB :)
<Werner> Indeed
<IgorPec> i only boot one device and didn't actually check anything
<Werner> http://ix.io/2qos for the 1+
<IgorPec> dt bits names could be changed or something
<IgorPec> any other board like H5?
<Werner> Have no boards here
<Werner> The only board I have for sunxi64 is the 1+
<Werner> Others are sunxi
<ArmbianTwitter> @RichCheneyAzure (Rich Cheney): @ThE_JacO @xoofx @armbian Never seen “passive egressiveness” before. Definitely using that. (18s ago)
<IgorPec> ok, anyway i planned to run upgrade on my test setupo
<Werner> I'll give the OPi1 a shot too to see if it is broken on sunxi too
<Werner> Hm...maybe :D
<Werner> Go big or go home
<IgorPec> haha, i will kick my kids out of the room and install racks :)
<Werner> Hahaha
<Werner> cpufreq still working on sunxi
<IgorPec> aha, probably its related to h6 only
<IgorPec> that is my suspicious
<IgorPec> h3 and h5 have the same files
<Werner> Makes sense
xec has quit [Remote host closed the connection]
xecuter has joined #armbian
dddddd has quit [Ping timeout: 260 seconds]
Net147 has quit [Quit: Quit]
Net147 has joined #armbian
Net147 has quit [Client Quit]
Net147 has joined #armbian
Net147 has joined #armbian
Net147 has quit [Changing host]
Net147 has quit [Remote host closed the connection]
Net147 has joined #armbian
Net147 has joined #armbian
Net147 has quit [Changing host]
sassinak-work has quit [Ping timeout: 246 seconds]
sassinak-work has joined #armbian
archetech has quit [Quit: Konversation terminated!]
prist has quit []
archetech has joined #armbian
archetech has quit [Quit: Konversation terminated!]
<ArmbianTwitter> @DieZuckerbude (Ben Zucker 🍰): @orangepixunlong #Orangepi one plus running @armbian based @debian #Bullseye with #Panfrost on #Linux mainline and #MATE desktop environment with #Compiz extension. Not very smooth at all but running. Nice. https://t.co/wSsWf9I0zX (30s ago)
<ArmbianTwitter> @armbian (armbian): RT @DieZuckerbude: @orangepixunlong #Orangepi one plus running @armbian based @debian #Bullseye with #Panfrost on #Linux mainline and #MATE… (14s ago)
<ArmbianTwitter> @RealDanct12 (Danct12): @DieZuckerbude @armbian @orangepixunlong @debian I kinda miss Compiz when I first try it in a VM, that's like 8-9 years ago. Nowadays Compiz is quite dead, or is it? (15s ago)
Hokedli has joined #armbian
<ArmbianTwitter> @DieZuckerbude (Ben Zucker 🍰): @RealDanct12 @armbian @orangepixunlong @debian Mate has an integration package: https://t.co/wxiejlxfN4 (11s ago)
archetech has joined #armbian
xecuter has quit [Remote host closed the connection]
xecuter has joined #armbian
xecuter has quit [Remote host closed the connection]
xecuter has joined #armbian
xecuter has quit [Remote host closed the connection]
Hokedli has quit [Quit: Konversation terminated!]
sassinak-work has quit [Ping timeout: 240 seconds]
sassinak-work has joined #armbian
xecuter has joined #armbian
dddddd has joined #armbian
awf_sbrt has quit [Read error: Connection reset by peer]
<ArmbianTwitter> @gnuhealth (GNU Health): Proud to announce #GNUHealth embedded for @Olimex A20 - LIME2 & @armbian! LibreSoftware + LibreHardware 💚 #GNUHealthEmbedded . Stay tuned for the images! https://t.co/kuQ0bnkOLu (20s ago)
chewitt_ has joined #armbian
chewitt has quit [Ping timeout: 260 seconds]
chewitt_ has quit [Ping timeout: 260 seconds]
DaRock has quit [Ping timeout: 264 seconds]
chewitt has joined #armbian
<lanefu> Tonymac32: i noticed our menson64 current, and dev patches are different despite same kernel
<lanefu> tried dev kernel to see if it would fix my res problem, but no dice.. building a unpatched kernel now
<IgorPec> lanefu: my N2 does not always detect resolution
<lanefu> hardsetting does nothing for me in boot.ini
<IgorPec> so its not just yours
<lanefu> here's what Xsee... it gets the resolutions, but not refresh http://ix.io/2qpM
<lanefu> s/Xsee/Xorg Sees at startup/
<ArmbianHelper> lanefu meant to say: here's what Xorg Sees at startup... it gets the resolutions, but not refresh http://ix.io/2qpM
<IgorPec> yeah, but screen is not alright already at boot
<IgorPec> when its fuc* up
gediz0x5391 has joined #armbian
<ArmbianTwitter> @armbian (armbian): RT @gnuhealth: Proud to announce #GNUHealth embedded for @Olimex A20 - LIME2 & @armbian! LibreSoftware + LibreHardware 💚 #GNUHealthEmbedde… (11s ago)
<lanefu> Tonymac32: tried test again on odroid with kernel 5.7.. same http://ix.io/2qpX
IgorPec has quit [Remote host closed the connection]
IgorPec has joined #armbian
IgorPec has joined #armbian
redhoticaro has joined #armbian
<redhoticaro> good evening
<redhoticaro> can I send a question?
<archetech> you dont ask bad etiqutte
<redhoticaro> excuse me?
<archetech> you're asking for personal attn
<redhoticaro> attn?
<archetech> attention
<redhoticaro> well...I just wanna try to solve a problem with my Odroid c2
<archetech> state your board and question for it
<archetech> ok odroid c2
<redhoticaro> yes ok..the problem is that desktop enviroinment want start up after updating
<redhoticaro> wont start up
<archetech> which armbian image ?
<redhoticaro> i Can still log via ssh so I can use terminal
<archetech> ubu deb ?
<redhoticaro> armbian 5.6.15 meson 64
<archetech> ok now wait for someone to help
<redhoticaro> ok
<lanefu> look in /var/log/Xorg.0.log for errors
<IgorPec> before going any further, armbianmonitor -u
<redhoticaro> OK
<redhoticaro> says : System diagnosis information will now be uploaded to invalid message format Please post the url in the forum where you've been asked for
<redhoticaro> by the way..in the log file there are some errors
<redhoticaro> (EE) Screen1 deleted because of no matching config section.
<redhoticaro> (EE) Screen2 deleted because of no matching config section.
<IgorPec> url ?
<redhoticaro> (EE) Fatal server error: (EE) Cannot run in framebuffer mode. Pleasde specify bus IDs for all framebuffer devices
<IgorPec> until we don't see logs which we want to see, its pointless
<redhoticaro> IgorPec yes
<IgorPec> linux is a work of 10000 people
<lanefu> redhoticaro: did the armbinamonitor -u return a ix.io url?
<redhoticaro> no
<IgorPec> but you have a network on that computer?
<redhoticaro> IgorPec I've tried to copy and past the log file from terminal but I don't know how
<redhoticaro> IgorPec yes
<redhoticaro> is my home network
<IgorPec> if you can't provide logs with our tool, its more wrong that just a desktop not getting up. probably
<IgorPec> rather start from a clean image.
<redhoticaro> which is your tool?
<lanefu> redhoticaro: try ` sudo armbianmonitor -u`
<redhoticaro> I did it but the command return this : " System diagnosis information will now be uploaded to invalid message format Please post the url in the forum where you've been asked for"
<lanefu> redhoticaro: what shell are you using
<lanefu> but yeah stuff sounds prettty broken :(
<redhoticaro> typing ctrl+alt+f1
<IgorPec> it seems the service is down
<redhoticaro> ok I tried again
<redhoticaro> now it shows an url
<IgorPec> great
<redhoticaro> the url is http://ix.io/2qqf
<Tonymac32> lanefu yeah, display has always been a problem on amlogic, the clock system is split between kernel, uboot and the infamous "blob". Now, why it doesn't like the CEC info I really can't say.
<Tonymac32> lanefu for XU4 stuff, that is the Hardkernel kernel doing the same?
<lanefu> Tonymac32: 5.4x is hardkernel... 5.7 is mainline
<IgorPec> reboticaro: how is the symptom?
<redhoticaro> well is that the desktop environment doesn't bootup
<redhoticaro> works only the shell typing ctrl+alt+f1
<lanefu> redhoticaro: had you done in customizations to the desktop config for 3d or acceleration
<redhoticaro> nope
<redhoticaro> just run apt update and apt upgrade and after that...bye bye
<IgorPec> but the desktop is black screen or desktop without menues?
<IgorPec> its bionic desktop where mesa drivers are broken
<redhoticaro> black screen with only in the top-left the white dash
<lanefu> Tonymac32: you think this is more of a u-boot problem than a kernel problem
<redhoticaro> also tried apt -f install and just rescued teamviewer
<redhoticaro> wich shows black screen when loggin in as well=(
<Tonymac32> earlier I said CEC, I meant EDID
<Tonymac32> but yes, it could be
<IgorPec> i'll test it on hardware. lately i haven't done much tests with bionic
<Tonymac32> we need to collect all of the info we can and compile it in one place and let baylibre know, we'll need the EDID/etc dumped per monitor and board with issue
<Tonymac32> lanefu ^
gediz0x5391 has quit [Quit: Leaving.]
<Tonymac32> I've had issues with every kernel including 4.19
gediz0x539 has joined #armbian
<redhoticaro> is there a fix or not?
<IgorPec> we don't know
<redhoticaro> so maybe I should install a new image right?
<IgorPec> wait few minutes
<redhoticaro> ok
<redhoticaro> thanks
<IgorPec> i will see if i can reproduce. at least something
<redhoticaro> ok
<IgorPec> yes, bionic desktop is broken
<IgorPec> but now i have to see if its a kernel or ubuntu problem
stid has joined #armbian
<stid> I'm running armbian bullseye on NanoPi M4V2. The source list looks like:
<stid> deb http://beta.armbian.com bullseye main bullseye-utils bullseye-desktop
<stid> And after upgrading today I found that both 'linux-u-boot-nanopim4v2-current' and 'armbin-firmware' are missing. Is this expected?
dddddd_ has joined #armbian
<IgorPec> stid: no, this means something is wrong
<IgorPec> you upgrade how?
<stid> I'm still keeping those two packages but they don't seem to be listed in the package list
<IgorPec> redhoticaro: get Ubuntu Focal, Bionic is simply broken
<stid> I did the upgrade through 'aptitude' by pressing 'u' and 'U', then 'g'. That's it
dddddd__ has joined #armbian
<IgorPec> broken upstream and that is low priority to fix
dddddd has quit [Ping timeout: 264 seconds]
<redhoticaro> IgorPec i have to reinstall image from the beginning
<stid> I guess it would result the same if I run 'apt-get update' and 'apt-get upgrade'
<redhoticaro> ?
<IgorPec> redhoticare: yes. or you fix problems with Ubuntu
<IgorPec> Ubuntu Bionic desktop is fragile ... one thing you try
<stid> IgorPec I checked the packages at beta.armbian.com, and both packages are missing too
<IgorPec> add Option "AccelMethod" "none" to /etc/X11/xorg.conf.d/01-armbian-defaults.conf
<IgorPec> and reboot
<IgorPec> stid: yes, i see. i have to dig-in and understand why
<redhoticaro> IgorPec you suggest ubuntu not armbian?
dddddd_ has quit [Ping timeout: 260 seconds]
<stid> IgorPec Thanks for looking into this
<IgorPec> redhoticaro: ubuntu will not boot at all
<IgorPec> redhoticaro: Armbian comes with many variants and yours is broken
<redhoticaro> ok
<IgorPec> its Ubuntu Bionic based
<IgorPec> i just check Ubuntu Focal based and desktop works. but Bionic might be fixed if you add that to the XORG conf
<IgorPec> you can also upgrade to Ubuntu 20.04 ... but its yet another adventure
<IgorPec> check forums. some folks succedded
<redhoticaro> at thjis point i prefer to start from begininnig
<redhoticaro> thank you
<IgorPec> u welcome
redhoticaro has quit [Remote host closed the connection]
dddddd__ is now known as dddddd
<stid> IgorPec The packages seem to be there at http://apt.armbian.com. Should I switch to this mirror instead? I wonder what is the difference between beta.armbian.com and apt.armbian.com
<IgorPec> beta is for development
<IgorPec> as a name suggests. can be broken
<IgorPec> here things are updated when changes are made to the code in any way, while apt repository is more tested
<lanefu> man we need recruit a desktop team to worry about desktop config and upgrade testing on that frnt
<IgorPec> yeah. lets start with a forum topic ?
<IgorPec> aka announcement to seek out for a desktop ceo
<IgorPec> :)
<fromport> is there a way i can force armbian to not rename ethernet like you can do on intel in /etc/default/grub GRUB_CMDLINE_LINUX="net.ifnames=0 biosdevname=0" ? i would like it to survive upgrades with newer kernels
<stid> IgorPec The issue is gone after I switch to 'apt.armbian.com'. Thanks anyway
<IgorPec> extraargs= is the kernel parameter on armbian . add it to /etc/armbianEnv.txt
<fromport> IgorPec: ty!
Hokedli has joined #armbian
stid has left #armbian [#armbian]
dddddd_ has joined #armbian
sunshavi has quit [Read error: Connection reset by peer]
dddddd has quit [Ping timeout: 260 seconds]
<lanefu> IgorPec: Desktop CEO.. i like it
archetech has quit [Quit: Konversation terminated!]
archetech has joined #armbian
plntyk has quit [Quit: Leaving]
sunshavi has joined #armbian
mpmc has quit [Quit: ZNC - https://znc.in]
mpmc has joined #armbian
Hokedli has quit [Quit: Konversation terminated!]
dddddd__ has joined #armbian
dddddd_ has quit [Ping timeout: 246 seconds]
dddddd_ has joined #armbian
dddddd__ has quit [Ping timeout: 260 seconds]
dddddd has joined #armbian
dddddd__ has joined #armbian
dddddd_ has quit [Ping timeout: 265 seconds]
dddddd has quit [Ping timeout: 264 seconds]
sassinak-work has quit [Ping timeout: 244 seconds]
sassinak-work has joined #armbian
xec has joined #armbian
xecuter has quit [Ping timeout: 260 seconds]
sassinak-work has quit [Ping timeout: 272 seconds]
sassinak-work has joined #armbian
archetech has quit [Quit: Leaving]
archetech has joined #armbian
gediz0x539 has quit [Quit: Leaving.]