lanefu changed the topic of #armbian-rockchip to: Armbian - Linux for ARM development boards | Rockchip SoC | www.armbian.com | This channel is relayed to the equivalent Discord channel | this channel is logged
<Armbian-Discord> <R​ichN> does you board have emmc ?
<Armbian-Discord> <R​ichN> as it is not the script looks for emmc and nvme
<Armbian-Discord> <R​ichN> as it has to put /boot on emmc for booting and /root on nvme
<Armbian-Discord> <R​ichN> else it will fail unles you have a spi-nore flash that has the bl on it
<Armbian-Discord> <j​oekhoobyar> Already dissatisfied with the 2nd adapter. It did a boot loop for 25 seconds before finally booting.
<Armbian-Discord> <j​oekhoobyar> I'm gonna do the stress test anyway, but I imagine I will return these
<Armbian-Discord> <j​oekhoobyar> I guess i could be misinterpreting the 25 second delay - even though I saw the led tun off and back on again at least twice.
<Armbian-Discord> <j​oekhoobyar> First boot was on an A1 sdcard with 75/mbs seq write speed - now it's on eMMC root, NVME for the rest
<Armbian-Discord> <j​oekhoobyar> Okay, @Tonymac32 - I guess I may keep these after all. 🙂 Stress test seems to be fine. Will run it some more too
<Armbian-Discord> <j​oekhoobyar> ```
<Armbian-Discord> <j​oekhoobyar> # stress-ng --matrix 0 --tz -t 5m --times
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] dispatching hogs: 8 matrix
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] successful run completed in 300.00s (5 mins, 0.00 secs)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] matrix:
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] bigcore0-thermal 73.00 C (346.15 K)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] bigcore1-thermal 73.00 C (346.15 K)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] center-thermal 70.23 C (343.38 K)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] gpu-thermal 68.62 C (341.77 K)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] littlecore-thermal 69.12 C (342.27 K)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] npu-thermal 68.69 C (341.84 K)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] soc-thermal 68.91 C (342.06 K)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] for a 300.00s run time:
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] 2400.04s available CPU time
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] 2397.66s user time ( 99.90%)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] 0.03s system time ( 0.00%)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] 2397.69s total time ( 99.90%)
<Armbian-Discord> <j​oekhoobyar> stress-ng: info: [6317] load average: 7.96 5.59 2.68
Armbian-Discord has quit [Remote host closed the connection]
Armbian-Discord has joined #armbian-rockchip
<Armbian-Discord> <j​oekhoobyar> I have a fan on it - or it would have gotten hotter for sure
<Armbian-Discord> <j​oekhoobyar> only peripheral is the NVME drive and the lan
<Armbian-Discord> <j​oekhoobyar> a pretty meager test, I guess 😕
<Armbian-Discord> <T​onymac32> You can only do so much
<Armbian-Discord> <j​oekhoobyar> This board's role is just a k3s agent node
<Armbian-Discord> <j​oekhoobyar> with piraeus (linstor / drbd) storage
<Armbian-Discord> <j​oekhoobyar> so far i like piraeus better than longhorn, but I still need more time tweaking, learning, testing
<Armbian-Discord> <j​oekhoobyar> drbd has had a change bitmap since the 2000's, longhorn requires a full resync of the drive if the node goes down / up
<Armbian-Discord> <j​oekhoobyar> I really hate the radxa kernel - building the linstor drbd 9.1.x from source was utterly ridiculous
<Armbian-Discord> <j​oekhoobyar> their source deb is littered with binaries (cross-compiled on an amd64), as well as their headers deb
<Armbian-Discord> <j​oekhoobyar> do we do any repackaging / cleanup of their source / headers?
<Armbian-Discord> <j​oekhoobyar> because I could tack on to that if we are doing so
<Armbian-Discord> <j​oekhoobyar> anyway, across multiple armbian kernels as well as a debian kernel, this is the first kernel where linbit's drbd automatic patch detection fails to properly detect the kernel's settings
<Armbian-Discord> <j​oekhoobyar> likely due to the poor hygeine of the above
<Armbian-Discord> <T​onymac32> Their kernel is a 3.10 mingled with a lot of android crap, patched to bring it up to a facsimile of a modern kernel
<Armbian-Discord> <j​oekhoobyar> oh jesus, seriously? So, we gotta blame rockchip for that, right?
<Armbian-Discord> <j​oekhoobyar> what a mess
<Armbian-Discord> <T​onymac32> Yep
<Armbian-Discord> <j​oekhoobyar> I really wish I had kernel hacking chops, and I'd dive right in.
<Armbian-Discord> <j​oekhoobyar> But it would be a learning curve, and I'm still gonna be in "busy mode" with my startup for some time
<Armbian-Discord> <j​oekhoobyar> as much as I like to pretend to be superhero, I gotta at least have some semblance of limits on what I can get done 😆
<Armbian-Discord> <j​2p2> yes I have bootloader installed in spi nor flash, but setup i'm trying to use is /boot on sdcard and /root on nvme
<Armbian-Discord> <j​2p2> it worked a couple days ago until random kernel crashes started happening while using docker. Now it seems that kernel crashes while doing any sync with nvme (including trying to install to nvme like above)
<Armbian-Discord> <T​enkawa> This a Rock5?
<Armbian-Discord> <j​oekhoobyar> yes
<Armbian-Discord> <j​oekhoobyar> sorry, was recabling one of the Rock5(s) at that moment. trying to keep things organized
<Armbian-Discord> <j​oekhoobyar> still need to spend some time stressing drbd 9.1.x to ensure it works on the dirty 5.10.x kernel
<Armbian-Discord> <j​oekhoobyar> but it's late - and tomorrow is my first day back from vacation
<Armbian-Discord> <T​enkawa> The Rock5 (RK3588 in general) is definitely... unique
<Armbian-Discord> <j​oekhoobyar> the power of this SOC, I still think it's great. But the kernel, they really surprised me - before coming to Armbian, I had no idea how bad these things can be
<Armbian-Discord> <j​oekhoobyar> @Tenkawa - i know you beat on the 5.15.y for rock5 for a long time, was it just a losing battle in the end? or is it worth continuing to hack on?
<Armbian-Discord> <T​abalugach> I own 2 Rock5B, for my use case, it's working very well, very stable (running official rock5b image), now I start playing with Armbian (not so stable)
<Armbian-Discord> <T​abalugach> uptime 09:20:38 up 19 days
<Armbian-Discord> <T​abalugach> not sure why Rock5 sometimes failed to boot from nvme using armbian,
<Armbian-Discord> <T​abalugach> armbian reboot sometimes won't work, sometimes does, its pure random
<Armbian-Discord> <T​enkawa> I've only worked on 5.10 on it... as far as I know there is no 5.15 useable on it.
<Armbian-Discord> <T​enkawa> 5.10 I've pieced together parts to get it to a decently working 5.10.161 however it was very time consuming.
<Armbian-Discord> <T​abalugach> still playing with Armbian on rock 5b, after reboot or shutdown :
<Armbian-Discord> <T​abalugach> here i type reboot, and everything works well.. Any suggestions?
<Armbian-Discord> <R​ichN> any one running a ssd nvme or sata needs a heatsync and thermal pad tape by thermal grizzly
<Armbian-Discord> <R​ichN> if your ssd gets to hot it will also corrupt
<Armbian-Discord> <R​ichN> even if its in one of those external metal cases if there is nothing touching the ssd then it is getting no thermal release
<Armbian-Discord> <R​ichN> Taba I would take the sd card out and put it into antother system and fsck the sd card
<Armbian-Discord> <R​ichN> once dont there I woud clone it
<Armbian-Discord> <R​ichN> so that in the future you can just slid. in a backup sd card and boot