Age | Commit message (Collapse) | Author |
|
The current Graceful Restart time is 20 sec, but this value is not
enough long and can be exceeded unexpectedly while waiting for
re-establishing peer connections. Then the stale paths on a receiving
router are dropped before the paths successfully synced, and some test
cases on "graceful_restart_test.py" will be passed unexpectedly.
This patch extends the Graceful Restart time to 30 sec and introduces
the constants for Graceful Restart time and Long-Lived Graceful Restart
time.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Quagga seems to advertise routes to neighbors by intervals of 30 seconds
(eBGP) or 5 seconds (iBGP) by default. Because of this delay, timer in
"bgp_confederation_test.py", 60 seconds, can be exceeded when the UPDATE
messages need to go through multi ASs.
This patch fixes to use minimum advertisement-interval and reduces the
time of test cases.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
For the stability of some scenario tests, this patch fixes to wait for
GoBGP starting up after restarting GoBGP daemon. Currently, there are
cases that CLI command can be called without checking whether GoBGP
daemon is surely started or not.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
GoBGPContainer.graceful_restart() does not "restart" GoBGP daemon, but
just stops (kills) GoBGP daemon. Then this patch renames
graceful_restart() to stop_gobgp() to avoid confusion.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
"exabgpcli" supports to show adj-RIB-in and adj-RIB-out, this patch
enables ExaBGPContainer to get these RIBs information.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Currently, to advertise or withdraw routes with ExaBGPContainer, we need
to configure static routes via config file and restart ExaBGP daemon.
In other words, we can NOT send withdrawing advertisement with
ExaBGPContainer. Also, restating ExaBGP daemon frequently can make
scenario test unstable, and it should be avoided.
On the other hand, with ExaBGP version 4.0.5 or later(*), we can
advertise or withdraw routes using "exabgpcli" without restating ExaBGP
daemon (or writing application which calls ExaBGP's APIs).
This patch fixes to use "exabgpcli" and reduces the number of restating
ExaBGP daemon.
Note: According to this change, adding routes into ExaBGP should be
called after adding neighbor.
(*): "exabgpcli" is introduced at version 4.0.2, but has some bugs
related to Python 3 compatibility and FlowSpec rules combinations, then
we need to use version 4.0.5 or later.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Some times, on Travis-CI, some test cases fail unexpectedly in checking
paths in RIBs due to advertisements are not yet received from other
routers.
Currently, in order to avoid this unexpected result, "time.sleep" is
inserted after adding new routes, but it is not enough.
This patch introduces a new function to enable to do assertion several
times and avoid failure with the first assertion.
Note: This patch does not introduces this change into all test cases, do
only into some test cases which fail relatively frequently.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
With the current implementation, when adding routes on GoBGPContainer,
we re-configure and reload daemon, then add routes using CLI command.
Differ from QuaggaBGPContainer, when deletion of routes, reloading is
not required, but if reloading daemon frequently, we need to wait for
the daemon restarting, it causes testing time longer and unstable.
This patch enables to add routes without reloading daemon.
Note: According to this change, adding routes should be called after
starting daemons.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Currently, to advertise or withdraw routes on QuaggaBGPContainer, we
need to configure static routes in configure file and restart Quagga
daemons.
In other words, we can NOT send withdrawing advertisement from Quagga
side.
Also, restating Quagga daemons frequently can make scenario test
unstable, and it should be avoided.
This patch fixes to use "vtysh" and reduces the number of restating
Quagga daemons.
Note: According to this change, adding routes on Quagga should be called
after starting daemons.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Currently, scenario test library does not wait for the daemons boot in
containers at starting up or after sending SIGHUP for reloading new
configurations.
So some test cases (e.g., graceful_restart_test.py) fails occasionally
to insert static routes into GoBGP after reloading new configurations.
This patch fixes to wait for the daemons boot and improves the stability
of scenario tests.
Note: This patch does not introduce these improvements to
BagpipeContainer and ExaBGPContainer, because;
- Docker image for BaGPipe, which used in scenario test library, is too
old and BagpipeContainer does not seem to be used in the current test
cases.
- The version of ExaBGP in "osrg/exabgp" Docker image has no way to ask
the daemon status other than "ps" command ("ps" is already done, but
not enough). If "exabgpcli" is available on ExaBGP container, which is
required the next release of 4.0.2 or 3.4.21, we can use it to check
whether ExaBGP is started up or not.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Note: This implementation supposes the YABGP version 0.4.0.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Currently, in case of the FlowSpec routes, "prefix" value which stored
on each container instance is just a keyword and not valid NLRI
arguments for adding or deleting routes.
So when deleting routes with "route" keyword, it is required to
construct the NLRI arguments using "matchs" value.
This patch enables to construct the valid NLIR arguments as "prefix"
value when adding the FlowSpec routes and enables to delete the FlowSpec
routes with "route" keyword.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
With this patch, it is not required to enable IPv6 forwarding by using
"sysctl" explicitly in Quagga or GoBGP(Zebra enabled) containers.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
When creating a new Docker network (Linux Bridge), some routes to its
subnet will be installed in to the container host's routing table.
This routes enable containers to communicate each other through the
container host's rouging table, then causes the unexpected success of
scenario tests (e.g., ping between containers).
This patch removes the unexpected routes from the container host and
avoid this problem.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
|
|
For Unnumbered BGP, neighbor AS number isn't needed
to be specified, so 'peer-as' should not be specified
in the scenario test for Unnumbered BGP.
However, currently 'peer-as' is always specified
in test/lib/gobgp.py
This commit solves this by not specifying `peer-as`
if 'neighbor-interface' is configured.
Signed-off-by: Satoshi Fujimoto <satoshi.fujimoto7@gmail.com>
|
|
This commit introduces 'bgp_config' attribute for
GoBGP/Quagga containers to configure complex settings.
Adding to this, this commit enables to configure
BGP confederation settings to GoBGP/Quagga conrainers.
Signed-off-by: Satoshi Fujimoto <satoshi.fujimoto7@gmail.com>
|
|
For Quagga containers, currently, when we use get_global_rib()
without specifying prefix, we cannot get AS_PATH and other attributes.
And when we use get_global_rib() with prefix,
we can get only one path even if there is multiple paths for a destination.
This commit fixes these problems by refining test/lib/quagga.py.
Signed-off-by: Satoshi Fujimoto <satoshi.fujimoto7@gmail.com>
|
|
Signed-off-by: Satoshi Fujimoto <satoshi.fujimoto7@gmail.com>
|
|
Currently, the build steps of "go get" and removing the downloaded
directory are not effective steps.
For example, "go get" for downloading dependencies can be replaced
with "dep ensure".
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Currently, daemons of Quagga will fails to output their logs, because
those daemons (quagga user) does not have the permission to access to
the GoBGP config directory (owned by root user).
This patch prepares the volume for quagga user and enables daemons of
Quagga to output logs.
With this patch, those logs will be output into
"TEST_BASE_DIR/TEST_PREFIX/<CONTAINER_NAME>/quagga/" directory.
(default: /tmp/gobgp/<CONTAINER_NAME>/quagga/")
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
libraries are installed into vendor directory so we can't build go code outside.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
|
|
This patch enables GoBGPContainer to get community and identifier for
each route via get_global_rib() and get_local_rib() methods.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
This patch enables GoBGPContainer to add routes with "identifier" and
"community" via add_route() method.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Currently, del_route() with "identifier=None" deletes all routes even if
some installed routes have the valid identifier.
This patch fixes to delete only route which exactly matches against the
given identifier.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Signed-off-by: ISHIDA Wataru <ishida.wataru@lab.ntt.co.jp>
|
|
`go get -d github.com/osrg/gobgp/...` could fail (e.g., an imported
library has gone). After that, we build a docker image from modified
source code.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
|
|
we use the term replace-peer-as instead of as-override
since openconfig is using it.
cli
```
$ gobgp n add <neighbor-addr> as <asn> replace-peer-as
```
config
```
neighbor:
config:
peer-as: <asn>
neighbor-address: <neighbor-addr>
as-path-options:
config:
replace-peer-as: true
```
Signed-off-by: Wataru Ishida <ishida.wataru@lab.ntt.co.jp>
|
|
cli
```
$ gobgp n add <neighbor-addr> as <asn> remove-private-as (all|replace)
```
config
```
neighbor:
config:
peer-as: <asn>
neighbor-address: <neighbor-addr>
remove-private-as: all
```
Signed-off-by: Wataru Ishida <ishida.wataru@lab.ntt.co.jp>
|
|
RFC4271 9.1.2 Phase 2: Route Selection
If the AS_PATH attribute of a BGP route contains an AS loop, the BGP
route should be excluded from the Phase 2 decision function. AS loop
detection is done by scanning the full AS path (as specified in the
AS_PATH attribute), and checking that the autonomous system number of
the local system does not appear in the AS path. Operations of a BGP
speaker that is configured to accept routes with its own autonomous
system number in the AS path are outside the scope of this document.
Also this commit adds support for allow-own-as option to relax this.
Signed-off-by: ISHIDA Wataru <ishida.wataru@lab.ntt.co.jp>
|
|
fix the regression of commit e429c1aaf4dd85459dfafc1a5efac9d4778370f6
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
|
|
use cache wisely.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
|
|
This reverts commit 47ab84e34caeb7c89b26271bf84959011bc8ed19.
using the same binaries on a host and inside a container isn't always
feasible (by default, not statically linked binary).
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
|
|
Currently, to reflect the modification of the local source code or to
switch the base image for GoBGP container, it is required to re-build
the GoBGP container image, and this take a long time to test or debug
with the scenario tests.
This patch fixes to use the local GoBGP executables (gobgp and gobgpd)
in container, and enables to reflect the changes without rebuidling
container image.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Signed-off-by: ISHIDA Wataru <ishida.wataru@lab.ntt.co.jp>
|
|
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
With the newer version of Quagga, "en" command will be regarded as an
"Ambiguous command", because there are two commands "enable" and "end".
This patch replaces "en" command with "enable" command, and fixes this
problem.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
To enable to use "noseplugin" as the Python module from the absolute
import scope in scenario_test, this patch moves noseplugin.py to
test/lib module directory.
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
To avoid the intractive prompt before removals, this patch adds
"--force" option to remove directories.
Also, fixes "--recursive" to "--force" option when removing xml
files (not directories).
Signed-off-by: IWASE Yusuke <iwase.yusuke0@gmail.com>
|
|
When a path is rejected by in-policy, the prefix must be withdrawn
since it might already exist in the rib.
Signed-off-by: ISHIDA Wataru <ishida.wataru@lab.ntt.co.jp>
|
|
Signed-off-by: Wataru Ishida <ishida.wataru@lab.ntt.co.jp>
|
|
Signed-off-by: Wataru Ishida <ishida.wataru@lab.ntt.co.jp>
|
|
policy can have multiple statements. In each statement, when condition
is matched, route-action is finally evaluated after actions with mods.
When route-action is 'none', we contine to next statement if it exists.
When route-action is 'accept' or 'reject', we stop proceeding.
This patch fixes a bug that route-action can't be set to 'none' which
means route-action always be set to 'accept' or 'reject' and can't proceed
to the next statement.
Signed-off-by: Wataru Ishida <ishida.wataru@lab.ntt.co.jp>
|