This requires backporting two upstream commits to avoid a segfault
due to the /etc/iproute2/rt_addrprotos.d and
/usr/share/iproute2/rt_addrprotos.d directories not existing on OpenWrt,
and the following compile error:
In file included from /home/stijn/Development/OpenWrt/openwrt/staging_dir/toolchain-powerpc64_e5500_gcc-13.3.0_musl/include/net/ethernet.h:10,
from vrrp.c:44:
/home/stijn/Development/OpenWrt/openwrt/staging_dir/toolchain-powerpc64_e5500_gcc-13.3.0_musl/include/netinet/if_ether.h:115:8: error: redefinition of 'struct ethhdr'
115 | struct ethhdr {
| ^~~~~~
In file included from vrrp.c:43:
/home/stijn/Development/OpenWrt/openwrt/staging_dir/toolchain-powerpc64_e5500_gcc-13.3.0_musl/include/linux/if_ether.h:173:8: note: originally defined here
173 | struct ethhdr {
| ^~~~~~
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
If an interface that is being used (or tracked) by a VRRP instance goes to
down state, the VRRP instance(s) will, by default, immediately transition to
FAULT state, and when all relevant interfaces are back up again the VRRP
instance(s) will immediately transition to BACKUP state.
This can cause problems if interfaces are bouncing, and so delays can be
specified between the interface state change and the transition to
FAULT/BACKUP state. If the interface returns to its original state before
the delay expires, no associated VRRP instance state transition will occur.
New uci section 'interface_up_down_delay':
config interface_up_down_delays
option device <device>
option down_delay <number in seconds>
option up_delay <number in seconds>
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
This commit adds support for folder synchronization.
An example is in the 810-files file where it is explained how to
synchronize all hotplug files of keepalived.
Signed-off-by: Francesco Benini <francy.benini@gmail.com>
The hotplug scripts are called with every state change. When called, the
scripts are processed under '/etc/hotplug.d/keepalived'. This change adds
the functionality that the last state change of the keepalived can be
queried via the ubus.
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
This section was renamed some time ago. Although this would be fixed by
the uci-default migration script, we should set this correctly in the
example config file right away.
Signed-off-by: Martin Schiller <ms@dev.tdt.de>
When "set_reload_if_sync" is set, the service is reloaded when the
configuration changes. For dnsmasq this means that the service, if
stopped, is started, and we don't want this in the backup node.
Signed-off-by: Francesco Benini <francy.benini@gmail.com>
When "set_reload_if_sync" is not set in the hotplug script, the service
is not expected to reload. That is not true because even if not set, the
value is set to the default 1 (reload active) or equals the parameter
set when "keepalived_hotplug" is called.
The default behavior should be:
- Reload if set_reload_if_sync is called
- NOT reload if set_reload_if_sync is NOT called
A similar fix is ported to "set_update_target".
Signed-off-by: Francesco Benini <francy.benini@gmail.com>
Some init.d scripts like firewall and sqm do not return the actual state
of the service if called with "running" parameter. This result in the
init script called with "start" parameter and the service may not load
the new configuration. Firewall init script is one of this
An option is added in order to skip the "running" check for the service.
Signed-off-by: Francesco Benini <francy.benini@gmail.com>
The 'luci-app-keepalived' uses the status json output to parse this
information for the status page. The problem is that when the LuCI
status page is open in the browser, the query is logged every 3 second into
the syslog. This is not needed and can therefore be removed.
This patch was already merged upstream:
6cce75f4eb
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
This is not supported by keepalived uci configuration. If a script
should be called by a notify event, then the script must be placed under
the directory '/etc/hotplug.d/keepalived'.
Reported-by: Stijn Tintel <stijn@linux-ipv6.be>
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
The keepalived does support script call handling on start and stop.
All scripts located under '/etc/hotplug.d/keepalived' gets now called
with the env ACTION set to startup or shutdown. The script that want to
get called on this keepalived events could evalutate this env to run on
startup or shutdown.
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
Automatically compute and substitute current values for all
$(AUTORELEASE) instances as this feature is deprecated and shouldn't be
used.
The following temporary change was made to the core:
diff --git a/rules.mk b/rules.mk
index 57d7995d4fa8..f16367de87a8 100644
--- a/rules.mk
+++ b/rules.mk
@@ -429,7 +429,7 @@ endef
abi_version_str = $(subst -,,$(subst _,,$(subst .,,$(1))))
COMMITCOUNT = $(if $(DUMP),0,$(call commitcount))
-AUTORELEASE = $(if $(DUMP),0,$(call commitcount,1))
+AUTORELEASE = $(if $(DUMP),0,$(shell sed -i "s/\$$(AUTORELEASE)/$(call commitcount,1)/" $(CURDIR)/Makefile))
all:
FORCE: ;
And this command used to fix affected packages:
for i in $(cd feeds/packages; git grep -l PKG_RELEASE:=.*AUTORELEASE | \
sed 's^.*/\([^/]*\)/Makefile^\1^';);
do
make package/$i/download
done
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
So that the busybox configuration does not have to be adapted, the
dependency has been changed to coreutils-timeout, which provides the
same functionality.
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
add new package keepalived-sync to synchronize files and data
between master and backup node. The master node uses SSH over rsync
to send and the backup node will use inotifywatch to watch received files.
The master node can track rsync.sh script to send configuration file on
a backup node based on the vrrp_script configuration of the same script.
The backup node will have a keepalived-inotify service, which would watch
for newly received files and it would call hotplug event. Each service
can keep its respective script under the keepalived hotplug directory and
executes commands to stop, start service or update any config in real-time.
Whenever a switchover will happen, the backup node would have the latest
config and data files from the master node.
Hotplug events can be used to apply config when files are received.
Signed-off-by: Jaymin Patel <jem.patel@gmail.com>
- enable json by default to generate json stats
- add rpc to generate json status
- add kmod-nf-ipvs dependencies for virtual servers
- set default vip labels on virtual interfaces
- set process name for keepalived child processes
Signed-off-by: Jaymin Patel <jem.patel@gmail.com>
The genhash binary is only built when IPVS is enabled, so make its
installation depend on IPVS being enabled.
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
Openwrt is a single user system. So keepalived is runnig as root.
If we add the config options `script_user root` and
`enabled_script_security' the following warnings are gone.
> local1.info Keepalived_vrrp[5382]: SECURITY VIOLATION - scripts are
being executed but script_security not enabled.
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
Backport of upstream staging commit to fix compilation issue with x86_64 as
ppc64le build hosts.
42c020aab6
This fixes#10642.
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
This fixes a runtime startup error on system which does not have a
toplevel runtime directory for the pid file. On openwrt the pid is
located at /var/run and not on /run. To fix that add a configure option to
move the pid location to /var/run.
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
This fixes a runtime startup error for system, which does not have enabled the
kernel config option CONFIG_PROC_EVENTS.
This workaround was published on github under the following URL.
https://github.com/acassen/keepalived/issues/1119
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
The CONTRIBUTING.md requests an (or multiple) SPDX identifier for GPL
licenses. But a lot of packages did use a different, non-SPDX style with a
"+" at the end instead of "-or-later".
Signed-off-by: Sven Eckelmann <sven@narfation.org>
