5 Likes
2 Comments
This code contains the #server with everything you need to create your own instance.
Transfer.sh currently supports the #s3 (Amazon S3), #gdrive (Google Drive), #storj (Storj) providers, and local file system (local).
✅ #Windows10 in #KVM gezwängt
✅ #Steuererklärung fertig - nettes Ergebnis
✅ erster eigener #S3 #Storage mit #minio und #podman aufgebaut
✅ erste sehr vielversprechende Tests mit #restic absolviert
✅ Erkältung eingefangen
Und ihr so?
#Linux #Manjaro #AMD #Ryzen #x360 #13ar0xxx #s2idle #s0ix #s3 #deepsleep #acpi
WARNING : This tutorial is specific to that exact computer model, but the general idea may help debug other systems.
So, to whom may find that tutorial useful …
Context :
I have a 2020 HP X360 Envy 13-ar0xxx (Ryzen 7 3700U based, nothing fancy but nice hardware) ; bios version F.25 dated 04/2022
The darn thing has never been able to sleep under Linux because HP screwed buyers by only allowing the s2idle method to put the computer into sleep mode, and that method is broken on AMD Ryzen based platforms. Work is being done to support s2idle in kernel, but at a glacially slow pace.
HP refuses to acknowledge the problem (unlike Dell or Lenovo) and locked the AMI Bios option to switch between S2idle and the more common S3 (deep sleep) method. Thankfully, hibernation has always worked and the SSD is fast enough it’s not too painful. But. But it’s MY computer and I’m not going to be told what I can, can’t or should do.
After months of trying every “legal” way to make it happen without success, I finally (thanks to a bout of COVID) took some days to stick my greasy fingers into the root cause of the problem : ACPI tables.
First you need to install the acpi-tools. With Manjaro, use snap :
https://snapcraft.io/install/acpi-tools/manjaro
- create a work directory and place yourself inside that directory
- run :
acpi-tools.acpidump -o tables
to create a file with the currently running tables then
acpi-tools.acpixtract -a tables
to extract the individual tables (we need both FACP and DSDT tables)
- FACP :
acpi-tools.iasl -d facp.dat to disassemble the FACP table
edit facp.dsl
increase the “Oem Revision” field at the top by 1
set the “Hardware Reduced (V5)” field to 0
acpi-tools.iasl -sa facp.dsl to assemble your modified table
- DSDT :
acpi-tools.iasl -e ssdt*.dat -d dsdt.dat
(if you don’t xref ssdt tables, you won’t be able to recompile dsdt)
edit dsdt.dsl
add 1 to the last parameter of the DefinitionBlock() line - that’s the OEM version number
search for XS3 ; replace with _S3 (one occurence)
acpi-tools.iasl -sa dsdt.dsl to assemble your modified table
- Make the following folder structure: payload/kernel/firmware/acpi/
In the acpi/ folder add your compiled facp.aml and dsdt.aml
In the payload/ folder, run:
find kernel | cpio -H newc --create > overrideacpi.img
- Move the resulting overrideacpi.img to your /boot folder
Edit /etc/default/grub
Add mem_sleep_default=deep as the last parameter of the line GRUB_CMDLINE_DEFAULT
(like so : ) GRUB_CMDLINE_DEFAULT="... mem_sleep_default=deep"
Create a new line just below like so :
GRUB_EARLY_INITRD_LINUX_CUSTOM="overrideacpi.img"
Save /etc/default/grub
Update grub as usual : sudo update-grub
- Finally reboot. Check in dmesg your modified ACPI tables are indeed loaded, and do a :
cat /sys/power/mem_sleep, it should now return :
s2idle [deep]
Try putting the computer to sleep and enjoy the sound of silence… and be amazed when it wakes up at the slightest touch of the space bar !
last words : I haven’t tested all the the scenarios where this could go wrong (plugging or unplugging AC power while the computer is suspended, for instance), but it seems to do the job so far. In S3 suspend mode, the computer draws less than 1% of the battery per hour.
This solution was heavily inspired by the work of Jordan Maris : https://gitlab.freedesktop.org/drm/amd/-/issues/1230#note_580057
Decided to spin up a local k3s cluster running on my (ARM64) laptop. Another interesting bit about the Docker environment is how easy it is to migrate configurations across platforms.
I'll add that spinning up a cluster in k3s is just running a single command per node; one for the master node and one for each of the server nodes. It's trivial to automate and completes in seconds.
Now I'm messing around with #ceph for managing high-availability #storage (filesystem and #s3) and #stolon for high-availability #postgres.
After upgrading to #postgres 14 I got strange errors whenever I tried to back up to a new bucket on my #S3 server in the basement. I ended up making a few other tweaks to the database (including upgrading it to Postgres 14.1), and the errors went away. Once again I have encrypted streaming offsite backups of the database, never more than ten minutes old, with daily checkpoints.
In another day or two I’ll delete the bucket that holds the ~150GB of Postgres 13 backups.
S4 is 100% compatible AWS S3 storage, accessed through Tor and distributed using IPFS.
Tor acts as a DNS and hides the physical location of the S4 server. IPFS acts as a CDN and will make your data permanently accessible and is impossible to take offline once it has been published. A sidecar docker container is provided to seamlessly proxy requests from your existing S3 code over Tor to S4.
https://github.com/anthonybudd/S4
#github #ipfs #s4 #nodejs #docker #tor #data #s3 #server #awssdk #minio
Hallo zusammen, ich bin #NeuHier. Ich interessiere mich für #ecovillages, #genossenschaften, #genossenschaftsgruendung, #genossenschaftsmodell, #globalvillage, #integralevolutionary, #integralevolutionarypractice, #integrallifepractice, #integraltransformativepractice, #managementsysteme, #managementsystems, #neweconomics, #neweconomy, #organisationsentwicklung, #organizational-development, #persönlichkeitsentwicklung und #s3 or #sociocracy30