pekkal
-
Posts
7 -
Joined
-
Last visited
Content Type
Forums
Store
Crowdfunding
Applications
Events
Raffles
Community Map
Posts posted by pekkal
-
-
On 9/1/2020 at 7:22 AM, gprovost said:
@fri.K That's very interesting info and not the first time NFS has been pointed as a possible root cause :
Have you tried some different NFS block sizes (wsize and rsize) settings on your client side ? or maybe reduce the number of NFS daemon ?
Because most probably under NFS load with default settings the system reach a unresponsive state and the hw watchdog kicks in and reset the system.
Check if watchdog service is running (systemctl status watchdog.service), if yes then you could also disable it to remove it from the equation during your troubleshooting.
@gprovost FYI: I've had zero issues past 6 months after starting umounting MacOS NFS mounts when no longer using them. When I forget to do this (which is often) & Mac goes to sleep a few times there will be tens of NFS daemons running on the Helios/Debian server. I do believe there is an issue related to NFS between MacOS & Debian. Issuing 'netstat | grep nfs' every now and then might help to confirm this.
-
@gprovost (how do I refer to a person/userid in this forum?)
re: "how to debug"
I've had no further crashes after the two incidents ≈10 days ago. With little to back this up I suspect the issue is related Mac OS NSF client implementation with Armbian server:
- the issues in my 10 months stable system started after I started working at home (due to the corona virus restrictions): I had my laptop NFS mounted to Helios for weeks
- the two crashes took place when I was not working: the Mac had gone to sleep. They did not appeared when Helios was under heavy load.
- Now I've (usually) unmounted the NFS disks after use: no issues past 10 days
With this: if the NFS hypothesis is true I would not find anything with the suggested steps. However, doing it with NFS shares to my Mac OS client going to sleep might (in)validate my suspicion? I believe MACs don't sleep all that well: they keep on waking up and checking things.
pekkal
-
@gprovost I get 12.58 V, obviously w/o any load.
-
How to debug?
After running since August with no issues, my Helios4 has now crashed twice during past week: the OLED remains visible without being updated, the fan remains quiet (unlike after shutdown), all services are down. After unplugging & plugging the power the system starts & all services are OK, but obviously the logs are gone. After reboot log.hdd has very little, dmesg shows no errors.
How can I move logs to permanent storage or what can I do otherwise post mortem ... or is it the time to rebuild the system on a new SD card?
Regards,
pekkal
-
On 8/23/2019 at 6:30 AM, gprovost said:
@customer001 Please send me private message with your details and order number.
@pekkal Well based on the picture I can see cracks on the OLED display (bottom right corner of the diplay), so you might have tighten the screw a bit too hard and bend the PCB which can have effect on how operate the display. You can also send me private message and see what we can do.
Thanks! Changing the display fixed the issue: tighten the screws with fingers only!
-
Helios4 OLED
I've just received my batch 3 Helios4 and gotten it running: moved drives, configured NFS, DLNA, SMB, FTP etc etc etc.
Works beautifully and silently (especially when the drives spin down!).
One issue though: the OLED display appears to be unreadable:
https://drive.google.com/file/d/17EH_zAGJaqeO0TbMkG_gp6ENy8TnFhya/view?usp=sharing
https://drive.google.com/file/d/1Y9MYaMlG0geIiBjJQx86xiZ21SF8EzOS/view?usp=sharing
Any ideas what went wrong ... the installation of HW and SW seemed easy enough with the wiki instructions...
Helios4 Support
in Marvell mvebu
Posted
Re: potential NFS related MacOS issues, just for info:
I've now had my Mac on all weekend, going to sleep maybe ten times: there are now 425 NFS connections to my Helios 4 / Debian Buster server, all allocated in the 'well known ports' (0-1023) area.
The impact on memory is minimal (20-30MB), but I wonder what happens when the below 1024 port area is exhausted? Instead of finding out, I'll umount the three mounts...