Releases: threefoldtech/zos
v0.3.4
V0.3.4 fixes a couple of bugs introduced in v0.3.3.
It also bring some retry to the upgrade daemon to make the upgrade procedure more robust.
Changelog
0-OS main daemons
capacityd
- Revert df02144.
This was required to allow ipv4 only node to still report their uptime and thus be eligible for farming.
identityd
- #813 Retry to update if it fails
provisiond
v0.3.4-rc1
Fix #766: Make sure counters don't underflow (#816) * Fix #766: Make sure counters don't underflow * Add counter decrement test
v0.3.3
V0.3.3 is out. The main feature of this release is the ability to configure the size and type of the root filesystem of the containers.
It also brings bug fixes and improvement in the node status. Node that are not fully healthy won't report to be online anymore.
Check the changelog for a full overview of all changes.
Enjoy :-)
Changelog
0-OS main daemons
capacityd
- #808 Fix a bug where SRU reported was 0
contd
- fix merging of environment variable coming from the flist and the reservation
identityd
networkd
- send network interfaces information as soon as possible in the boot process
- #790 Don't crash networkd when it fail to create ndmz. Instead of crashing networkd we now just retry forever with an exponential backoff strategy. This prevent networkd from ending up in a loop which creates a lot of request to the explorer
- #632 prevent capacityd and provisiond to start if networkd is not ready
provisiond
- #772 Add reporting of ndmz pub4 and pub6 network interfaces to the explorer
- fix container counters decrement
- #795 allow to configure size and disk type of containers
- #805 #802 skip the unsupported workload type from a reservation
- zdb client: add timeout to avoid locking provisiond on non responsive container
v0.3.2
v0.3.3-rc1
Fix reported 0 sru (#808)
v0.3.2-rc2
allow to configure size and disk type of containers (#795) * allow to configure size and disk type of containers fixes #793 * contd: wait for shim-log binary to be present on the system before starting * go mod tidy * use the configured disk type for flist backend filesystem
v0.3.2-rc1
prevent capacityd and provisiond to start if networkd is not ready fixes #632
v0.3.1
v0.3.0
The v0.3.0 is out. This version bring very little change feature wise but fixes 2 major bugs.
The first bug allow the explorer to scale way better and reduce the loads generated by the node by a huge factor.
The second bug fixes an issue that prevented to reserve a volume and mount it in a container in the same reservation.
Changelog
0-OS main daemons
provisiond
#764 Always on pull all reservation once at boot
#761 Fix attaching volume container
extra binaries
#756 Extract shim-logs to external repository and move it to bins-extra
v0.3.0-rc2
make sure we always at least poll once on the latest unprocessed id