Release Notes #

Setting up Linux Virtual Server (LVS) in combination with SMTP did not work due to missing iptables -m ipvs support. IPVS (IP Virtual Server) is used for Linux Virtual Server.

The iptables package has been updated to a higher version and now includes the match support of iptables. To make it work, you need the probe kernel module xt_ipvs. It is provided by the respective cluster-network-* package.

The wizards could be difficult to fill in as they provided minimal information about valid parameter values.

Wizards now validate parameter values on the fly. They also display extended documentation about all parameters which includes showing the default value for the parameter (if any).

The wizards in Hawk were able to make cluster configuration changes, but not configure the resources themselves.

In the new Hawk (package hawk2 ), the wizards include steps to install packages, set up default configuration files and ensure that requirements are met as part of applying the cluster configuration changes.

Tagging resources is a new feature in Pacemaker. Editing and viewing them has been possible using the command line, but not from the Hawk UI.

The new Hawk (package hawk2 ) enables creating and editing resource tags.

The history explorer is a very powerful feature, but the user interface needed improvement.

Reports now have a description page showing a list of important node events and resource events, making it easy to see what has happened at a glance. The interface has been redesigned to provide more feedback as to which transition is being viewed. Navigation is simplified by the buttons Previous Transition and Next Transition.

The step from using Hawk to using the command line tools can be a challenge.

In Hawk, there is now a screen Command Log that displays all the corresponding crm commands that Hawk runs behind the scene. This makes it easy to see the corresponding crm command to an action in the user interface.

The wizards in Hawk were not usable from the command line, and crmsh had its own form of wizards that were implemented differently than in Hawk.

The wizards are now shared between crmsh and Hawk. crmsh is able to apply the old XML wizards for Hawk, and Hawk is able to run the new configuration script style wizards the same as crmsh.

With huge devices, the initial full synchronization took a lot of time even if no real data has to be synchronized. Most of the time, the command drbdadm new-current-uuid --clear-bitmap <resource> can be used to simply skip it, then make sure the data is consistent on file system level.

For some corner cases, users still need to do a full synchronization. This will take a long time, especially in an environment with low network bandwidth.

We now supply another method to zero out devices on both sides instead of doing a full synchronization from one node to the other. This helps when the network speed becomes the bottleneck of the synchronization. After the exchange of packets to the other node, the device on both sides will start to zero out locally.

For more information, see the documentation at https://www.suse.com/documentation/sle-ha-12/book_sleha/data/sec_ha_drbd_configure.html#sec_ha_drbd_configure_init (https://www.suse.com/documentation/sle-ha-12/book_sleha/data/sec_ha_drbd_configure.html#sec_ha_drbd_configure_init).