v0.2.4-rc1

provisiond: fix the amount of HRU reported as reserved

v0.2.3

v0.2.3 is out.

This release contains mostly bug fixes and improvements. This release pave the way to the enabling of the payment flow that will enable to use TFT token to pay for reservation. This new feature should be finished and enable in 0.3.0

Changelog

0-OS main daemons

bootstrap

• #596 remove a log message with error level that was creating a lot of confusion at boot.

capacityd

• #679 avoid spamming explorer if it fails to push capacity

containerd

• #502 Add support to secret environment variables for containers
• #578 Container can now stream their logs to a remove redis endpoint

identityd

• 580 output clearer error message when farm ID is wrong

networkd

• #556 make sure a network resource is deleted before releasing the associated wg port
• Make sure that public config is not overridden during not boot
• #556 properly report reserved port for wireguard to the explorer
• #593 the mac address of the NIC are also reported to the explorer
• #583 limit the amount of address update send to the explorer to 10 every 10 minutes
• increased the interval between checks for new public interface configuration

provisiond

• #554 don't block provision engine when a wrongly formatted reservation comes in
• fix a typo that prevent to send reserved RU to the explorer
• add missing PublicIP6 conversion from jsx schema to zos types
• #560 prevent persisted resources to be leaked forever
• #588 gedis: allow to re-use connections from the pool
• #614 Fixes dangling files in provision cache
• #683 sort the workloads from a reservation before sending it to provisiond engine

storaged

• #563 use wipefs to clean up v1 disks
• #564 run partprobe after creating the filesystem. This improve detection of the disk on server with a lot of them.
• #556 when storaged fails to mount the cache disk. It puts the node in a low-cache mode which doesn't allow any reservation to be provisioned. This is the first step toward a way to notify farmers when a disk is dead and need to be replaced.
• #608 fix a but that was preventing some of the nodes to boot because the disks were behind a raid controller and smartcl could not find them.

zinit

Documentation

• some cleanup has been done in the k8s documentation
• a new documentation about how to configure the network of a farm running 0-OS v2
• creation of a new compatibility list. We invite any farmer to contribute to it, see https://forum.threefold.io/t/hardware-compatibility-database/304
• creation of an FAQ section in the documentation
• improve documentation around reservation flow and payment

Tooling

explorer

• The bcdb_mock has been renamed to explorer and is now the official explorer server on all 3 network (mainnet, testnet and devnet)
• unify the explorer client. There is now an official go client for the explorer

CI

• #483 run tools test suite in CI
• upgrade go version to 1.14

tfuser

• rework of the identity file used by tfuser works. Now all the information about your identity are stored and a single flag is required on both tfuser and tffarmer to configure the identity

v0.2.2-rc9

provisiond: sort the workloads from a reservation before sending it to

provisiond engine

fixes #683

v0.2.2-rc8-b

provisiond: sort the workloads from a reservation before sending it to

provisiond engine

fixes #683

v0.2.2-rc8-a

Merge branch 'hot_fix' into 0.2.2-rc8-a

v0.2.2-rc7

ensure public key is always send during node registration

fixes #649

v0.2.2-rc6

update main readme

v0.2.2-rc5

disable debug tty

v0.2.2-rc4

always return path for volatir directory

v0.2.2-rc3

always close http response body