Update CoreOS manually
This article has been created by an automatic translation software. You can view the article source here.
CoreOS by default includes a built-in system update (update_engine) who plan an audit of updates available at regular intervals, automatically installs and program a reboot for that they are effective.
However, it may be useful in some cases to search and manually install the updates available. We will see below how to upgrade a stable CoreOS.
We assume you just deploy a CoreOS instance and that you are connected to is SSH user "core".
We begin by checking the current version of CoreOS :
core@CoreOSnode ~ $ cat /etc/lsb-release
DISTRIB_ID=CoreOS
DISTRIB_RELEASE=633.1.0
DISTRIB_CODENAME="Red Dog"
DISTRIB_DESCRIPTION="CoreOS 633.1.0"
We are launching the installation of updates (the return has been voluntarily truncated ) :
core@CoreOSnode ~ $ sudo update_engine_client -update
[0622/105739:INFO:update_engine_client.cc(245)] Initiating update check and install.
[0622/105739:INFO:update_engine_client.cc(250)] Waiting for update to complete.
LAST_CHECKED_TIME=1434963460
PROGRESS=0.000000
CURRENT_OP=UPDATE_STATUS_UPDATE_AVAILABLE
NEW_VERSION=0.0.0.0
NEW_SIZE=137131931
LAST_CHECKED_TIME=1434963460
PROGRESS=0.000000
CURRENT_OP=UPDATE_STATUS_UPDATE_AVAILABLE
...
NEW_VERSION=0.0.0.0
NEW_SIZE=137131931
LAST_CHECKED_TIME=1434963460
PROGRESS=0.993617
CURRENT_OP=UPDATE_STATUS_DOWNLOADING
NEW_VERSION=0.0.0.0
NEW_SIZE=137131931
LAST_CHECKED_TIME=1434963460
PROGRESS=0.000000
CURRENT_OP=UPDATE_STATUS_FINALIZING
NEW_VERSION=0.0.0.0
NEW_SIZE=137131931
LAST_CHECKED_TIME=1434963460
PROGRESS=0.000000
CURRENT_OP=UPDATE_STATUS_FINALIZING
NEW_VERSION=0.0.0.0
NEW_SIZE=137131931
<!--T:8-->
Broadcast message from locksmithd at 2015-06-22 10:59:57.244742724 +0200 CEST:
System reboot in 5 minutes!
<!--T:9-->
LAST_CHECKED_TIME=1434963460
PROGRESS=0.000000
CURRENT_OP=UPDATE_STATUS_UPDATED_NEED_REBOOT
NEW_VERSION=0.0.0.0
NEW_SIZE=137131931
[0622/105959:INFO:update_engine_client.cc(193)] Update succeeded -- reboot needed.
core@CoreOSnode ~ $
At the end of the installation of updates, we are invited to restart our instance CoreOS is immediately ("sudo reboot") either automatically at the end of 5 minutes.
After our CoreOS instance restarts, we check the new version thereof :
core@CoreOSnode ~ $ cat /etc/lsb-release
DISTRIB_ID=CoreOS
DISTRIB_RELEASE=681.2.0
DISTRIB_CODENAME="Red Dog"
DISTRIB_DESCRIPTION="CoreOS 681.2.0"
We are now in the latest stable CoreOS (681.2.0 at the time of this FAQ ).
Different update strategies /reboot
Integrated CoreOS 4 stratégies de reboot (reboot-strategy), ces stratégies contrôlent la façon dont le reboot se produit. Ces 4 strategies are :
- best-effort
- The parameter "by default in CoreOS ". This setting determines whether the machine is part of a cluster. If DCE is started the machine is supposed to be part of a cluster and therefore uses the strategy "DCE-lock ". Otherwise, it is the strategy "reboot" which is used.
- DCE-Lock
- Avec cette stratégie, chaque machine verrouille le reboot avant d'autoriser celui-ci. Cela permet à une mise à jour d'être appliquée rapidement à un cluster without losing quorum in DCE. Indeed, the number of machines permitted to restart simultaneously is configurable via the command (by adapting the maximum number to your context ) :
core@CoreOSnode ~ $ locksmithctl set-max 2
Old-Max: 1
Max: 2
- reboot
- This strategy will restart the machine as soon as the updates are installed.
- off
- Avec cette stratégie, la machine attend une commande demandant explicitement le reboot après installation des mises à jour.
The strategy is defined in the file /etc/coreos/update.conf :
core@CoreOSnode ~ $ cat /etc/coreos/update.conf
REBOOT_STRATEGY=off
be in the cloud-config file used like this :
#cloud-config
coreos:
update:
reboot-strategy: best-effort
[[category:cloud priv��]]
Enable comment auto-refresher