On a local CI run. The test_ldp_topo1.py showed fail to converge
on r3. r3 has 2 neighbors but only 1 was up when we got to
further steps in the test suites.
Modify the neighbor checking to `know` how many neighbors
should be operational and continue looking for them until
they are up and running.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Previously, when a valgrind memleak was discovered, would cause a
catastrophic pytest failure. Now correctly fails the current pytest as
intended.
As a result of this fix --valgrind-memleaks now works in distributed
pytest mode as well.
Signed-off-by: Christian Hopps <chopps@labn.net>
Revert the accidental enabling of the optional memleak tests that came
with the large micronet changeset.
Signed-off-by: Christian Hopps <chopps@labn.net>
The nexthop group code is installing routes and nexthop groups
and immediately expecting zebra to have processed the results
as a result there is a situation when the CI system is under
intense load that the nexthop group might not have been processed.
Add a bit of code to allow the test to give FRR some time
to finish work before declaring it not working.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When the CI system is heavily loaded, we might see the following failures:
```
test failed at "test_config_timing/test_static_timing": assert 20.083204 <= 19.487716
```
Currently we allow each step to run 2 times slower than the initial
measurement. Let's allow them to run 3 times slower.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
On the first step, the test creates 10000 static routes. It passes 10000
to `get_ip_networks` and it generates 10000 /22 routes.
On the fourth step, the test tries to remove 5000 previously created
routes. It passes 5000 to `get_ip_networks` and here starts the problem.
Instead of generating 5000 /22 routes, it generates 5000 /21 routes. And
the whole step is a no-op, we constantly see the following logs:
```
% Refusing to remove a non-existent route
```
To consistently generate same routes, `get_ip_networks` must always use
the same prefix length.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
Our topotests send SIGBUS 2 seconds after a SIGTERM is
initiated. This is bad because under a heavily loaded
topotest system we may have a case where the system has
not had a chance to properly shut down the daemon.
Extend the time greatly before topotests send SIGBUS.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
This removes a giant `switch { }` block from lib/zclient.c and
harmonizes all zclient callback function types to be the same (some had
a subset of the args, some had a void return, now they all have
ZAPI_CALLBACK_ARGS and int return.)
Apart from getting rid of the giant switch, this is a minor security
benefit since the function pointers are now in a `const` array, so they
can't be overwritten by e.g. heap overflows for code execution anymore.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
*_anywhere(item) returns whether an item is on _any_ container. Only
available for unsorted containers for now.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This provides a "is this item on this list" check, which may or may not
be faster than using *_find() for the same purpose. (If the container
has no faster way of doing it, it falls back to using *_find().)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Even if it doesn't matter for an unit test in general, it hides actual
leaks in the code being tested. Fix so any leaks will be actual bugs.
(Currently there aren't any, yay.)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This script is failing occassionally in our upstream topotests.
Where it was changing route-maps and attempting to see if
summarization was working correctly. The problem was that
the code appeared to be attempting to add route-maps to
redistribution in ospf then modifying the route-maps behavior
to affect summarization as well as the metric type of that
summarization.
The problem is of course that ospf does not appear to modify
the summary routes metric-type when the components
of that summary change it's metric-type. So the test
is testing nothing. In addition the test had messed
up the usage of the route-map generation code and all
the generated config was in different sequence numbers
but route-map processing would never get to those
new sequence numbers because of how route-maps are processed.
Let's just remove this part of the test instead of trying
to unwind it into anything meaningfull
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Several tests used the route_map_create functionality
with `metric-type` but never bothered to add the
backend code to ensure it works correctly.
Add it in so it can be used.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
We have this pattern in this test:
# Let's kill the interface on rt2 and see what happens with the RIB and BFD on rt1
tgen.gears["rt2"].link_enable("eth-rt1", enabled=False)
# By default BFD provides a recovery time of 900ms plus jitter, so let's wait
# initial 2 seconds to let the CI not suffer.
topotest.sleep(2, 'Wait for BFD down notification')
router_compare_json_output(
"rt1", "show ip route ospf json", "step3/show_ip_route_rt2_down.ref", 1, 0
)
Under a heavy CI load, interface down events and then reacting to them may not actually
happen within 2 seconds. Allow some more grace time in the test to ensure that we
react to it in an appropriate manner.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
OSPF when it is deciding on whom it should elect for DR and backup
has a process that prioritizes network stabilty over the exact
same results of who is the DR / Backups.
Essentially if we have r1 ----- r2
Let's say r1 has a higher priority, but r2 comes up first, starts
sending hello packets and then decides that it is the DR. At some
point in time in the future, r1 comes up and then connects to r2
at that point it sees that r2 has elected itself DR and it keeps
it that way.
This is by design of the system. With our tight ospf timers as
well as high load being experienced on our test systems. There
exists a bunch of ospf tests that we cannot guarantee that a
consistent DR will be elected for the test. As such let's not
even pretend that we care a bunch and just look for `Full`.
If we care about `ordering` we need to spend more time getting
the tests to actually start routers, ensure that htey are up and
running in the right order so that priority can take place.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Fix a loop in the setup phase of isis_topo1_vrf: only configure
interfaces that each router actually has.
Signed-off-by: Mark Stapp <mstapp@nvidia.com>
Ensure GR helpers have received a Grace-LSA before killing the
ospfd/ospf6d process that is undergoing a graceful restart.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
There's no more difference between number-named and word-named access-lists.
This commit removes separate arguments for number-named ACLs from CLI.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
BGP LU will use implicit-null in more situations now; adjust
the original LU topotest to align with that. Node R2 uses
imp-null now, while R1 continues to allocate labels.
Signed-off-by: Mark Stapp <mstapp@nvidia.com>
Add a second BGP labelled-unicast (BGP-LU) test suite, with
an additional router and some additional tests.
Signed-off-by: Mark Stapp <mstapp@nvidia.com>
2 things:
a) Each test was setting up for graceful restart with calls to
`graceful-restart prepare ip[v6] ospf`, then sleeping for
3 or 5 seconds. Then killing the ospf process. Under heavy
load there is no guarantee that zebra has received/processed
this signal. Write some code to ensure that this happens
b) Tests are issuing commands in this order:
1) issue gr prepare command
2) kill router
3) <ensure routes were still installed in zebra>
4) start router
5) <ensure routes were stil installed in zebra>
Imagine that the system is under some load and there is
a small amount of time before step 5 happens. In this
case ospf could have come up and started neighbor relations
and also started installing routes. If zebra receives
a new route before step 5 is issued then the route could
be in a state where it is not installed, because it is
being sent to the kernel for installation. This would
fail the test because it would only look 1 time. This
is fixed by giving time on restart for the routes to
be in the installed state.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Any command that uses `peer_lookup_in_view` crashes when "vrf all" is
used, because bgp is NULL in this case.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
FRR should only ever use the appropriate THREAD_ON/THREAD_OFF
semantics. This is espacially true for the functions we
end up calling the thread for.
Signed-off-by: Donatas Abraitis <donatas.abraitis@gmail.com>
Remove references to the deprecated "CLI()" function; clean up
a couple of string escapes; make one test-case sensitive to
previous failures.
Signed-off-by: Mark Stapp <mstapp@nvidia.com>
Some tests had commented-out references to the old "CLI()"
function. Remove those so they're not confusing in the future,
and replace at least one with a comment that uses the
'mininet_cli()' function.
Signed-off-by: Mark Stapp <mstapp@nvidia.com>
* Add new debug directives for NSSA LSAs;
* Remove the "debug ospf6 gr helper" command since it doesn't make
sense for this test (not to mention it was renamed to "debug ospf6
graceful-restart");
* Migrate to the new interface-level command to enable OSPFv3 on
interfaces ("interface WORD area A.B.C.D" was deprecated).
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Document the `sleep` statement so people know that we are sleeping
because we are waiting for the BFD down notification. If we don't
sleep here it is possible that we get outdated `show` command results.
Signed-off-by: Rafael Zalamena <rzalamena@opensourcerouting.org>
Call the `show` commands less often to reduce the CPU pressure.
Also increase the wait time from 60 to 80 seconds to have spare room
for failures (4 times more). This is the latest measure wait time:
> INFO: topolog: 'router_json_cmp' succeeded after 20.08 seconds
Signed-off-by: Rafael Zalamena <rzalamena@opensourcerouting.org>
Reduce timers so we send hello packets more often and reduce dead
interval to converge faster.
Previous test wait amount:
> INFO: topolog: 'router_json_cmp' succeeded after 47.20 seconds
New test wait amount:
> INFO: topolog: 'router_json_cmp' succeeded after 20.08 seconds
Signed-off-by: Rafael Zalamena <rzalamena@opensourcerouting.org>
FRR should only ever use the appropriate THREAD_ON/THREAD_OFF
semantics. This is espacially true for the functions we
end up calling the thread for.
Signed-off-by: Donatas Abraitis <donatas.abraitis@gmail.com>
Add the "default-information-originate" option to the "area X nssa"
command. That option allows the origination of Type-7 default routes
on NSSA ABRs and ASBRs.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
The route created by the "default-information-originate" command
isn't a regular external route. As such, an NSSA ABR shouldn't
originate a corresponding Type-7 LSA for it (there's a separate
configuration knob to generate Type-7 default routes).
While here, fix a small issue in ospf6_asbr_redistribute_add()
where routes created by "default-information-originate" were being
displayed with an incorrect "unknown" type.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Fix wrong comparison since route->path.metric_type is always set
to either 1 or 2. The OSPF6_PATH_TYPE_EXTERNAL2 constant, whose
value is 4, refers to a route type so its usage was incorrect here.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Considering that both the GR helper mode and restarting mode can be
enabled at the same time, the "graceful-restart helper-only" command
can be a bit misleading since it implies that only the helper mode
is enabled. Rename the command to "graceful-restart helper enable"
to clarify what the command does.
Start a deprecation cycle of one year before removing the original
command
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Because vrf backend may be based on namespaces, each vrf can
use in the [16-(2^32-1)] range table identifier for daemons that
request it. Extend the table manager to be hosted by vrf.
That possibility is disabled in the case the vrf backend is vrflite.
In that case, all vrf context use the same table manager instance.
Add a configuration command to be able to configure the wished
range of tables to use. This is a solution that permits to give
chunks to bgp daemon when it works with bgp flowspec entries and
wants to use specific iptables that do not override vrf tables.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Considering that both the GR helper mode and restarting mode can be
enabled at the same time, the "graceful-restart helper-only" command
can be a bit misleading since it implies that only the helper mode
is enabled. Rename the command to "graceful-restart helper enable"
to clarify what the command does.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Issue #9535 describes how the export-list/import-list commands work
differently on ospfd and ospf6d.
In short:
* On ospfd, "area A.B.C.D export-list" filters which internal
routes an ABR exports to other areas. On ospf6d, instead, that
command filters which inter-area routes an ABR exports to the
configured area (which is quite counter-intuitive). In other words,
both commands do the same but in opposite directions.
* On ospfd, "area A.B.C.D import-list" filters which inter-area
routes an ABR imports into the configured area. On ospf6d, that
command filters which inter-area routes an interior router accepts.
* On both daemons, "area A.B.C.D filter-list prefix NAME <in|out>"
works exactly the same as import/export lists, but using prefix-lists
instead of ACLs.
The inconsistency on how those commands work is undesirable. This
PR proposes to adapt the ospf6d commands to behave like they do
in ospfd.
These changes are obviously backward incompatible and this PR doesn't
propose any mitigation strategy other than warning users about the
changes in the next release notes. Since these ospf6d commands are
undocumented and work in such a peculiar way, it's unlikely many
users will be affected (if any at all).
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
A bunch of tests have this pattern:
a) Install a new prefix into bgp
b) Run this loop:
foreach (router in topology) {
verify_bgp_rib(router)
}
This is to ensure that the prefix is actually disseminated.
The problem with this, of course, is that a wait of 2 seconds
for every item in that loop makes no sense. As that the initial
router verification of it's bgp rib will wait 2 seconds and
all the remaining bgp routers in the topology will have gotten
the data. So we end up waiting a bunch of extra time.
Remove the initial_wait time for verify_bgp_rib. Also
increase the failure wait time to 30 seconds. This is
to give a bigger window for bgp to send it's data for
our test systems that could be under heavy load. In the
normal case tests will never hit this.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Add a new topotest that features a topology with seven routers spread
across four OSPF areas:
* 1 backbone area;
* 1 regular non-backbone area (0.0.0.1);
* 1 stub area (0.0.0.2);
* 1 NSSA area (0.0.0.3).
All routers have both GR and GR helper functionality enabled in
the configuration. The test consists of restarting each router,
one at time, and checking that all forwarding planes (and LSDBs)
are kept intact during those restarts.
A successful run takes about three minutes to finish.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Compilation is warning that a memcpy is only copying
the first (sizeof pointer) into memory. This is not
what we really want. Although it does beg the question about
why this memcpy is needed( or what it is doing ). I'm going
to just fix the memcpy and call it a day.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
frrmod_load() attempts to dlopen() several possible paths
(constructed from its basename argument) until one succeeds.
Each dlopen() attempt may fail for a different reason, and
the important one might not be the last one. Example:
dlopen(a/foo): file not found
dlopen(b/foo): symbol "bar" missing
dlopen(c/foo): file not found
Previous code reported only the most recent error. Now frrmod_load()
describes each dlopen() failure.
Signed-off-by: G. Paul Ziemba <paulz@labn.net>
1. Optimized test: test_clear_pim_neighbors_and_mroute_p0 run time by clearing
mroute and verifying mroutes separately. Execution time is reduced from almots 10 mins
to ~220 sec.
Signed-off-by: Kuldeep Kashyap <kashyapk@vmware.com>
route_scale run is 500+ seconds. Break it up into
2 separate tests. This should reduce run time a slight
bit.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Fix issue of topotest failures with BGP status Connect or Idle
instead of the expected Active
Signed-off-by: Martin Winter <mwinter@opensourcerouting.org>
Modernize the test a bit, generate expected results rather than load from
file, and add a general json_cmp with retry function and use it.
Signed-off-by: Christian Hopps <chopps@labn.net>
- Update the template and documentation to use newer pytest fixutres for
setup and teardown, as well as skipping tests when the suite fails.
Signed-off-by: Christian Hopps <chopps@labn.net>
When looking for a implied host route it is not necessary
to add the `/32` to an ip route add. As such masks
will not be set in this case. Set the value of masks
to a known good value so that when the route installation
fails the test for it actually being there will tell you
that the route is not there -vs- complaining about mask
being uninited.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
There is no need to add calls to addKernelRoutes for
groups. They do not need to be routed via the
normal kernel methodology.
Tests run successfully with this change.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
- Fix xterm support to work, previously it mostly didn't, not it should
in all cases (i.e., single or dist mode).
- Catch when the user tries to use various window requiring topotests
features (e.g., --cli-on-error) but isn't running under supported
system (e.g., byobu/tmux/xterm), and fail the run with an explanation.
Signed-off-by: Christian Hopps <chopps@labn.net>
If the SRv6 locator is deleted in zebra, zclient(bgpd)
which allocates SIDs from the locator will update the
RIBs which use those SIDs and make them invalid.
This will cause the VPNv6 route to be withdrawn and
the VPN to stop.
If the SRv6 locator is added again, zclient(bgpd) will
allocate the SIDs from the locator again, and VPNv6
will be re-established.
This commit add a test case to confirm this.
Signed-off-by: Hiroki Shirokura <slank.dev@gmail.com>
Before this PR, in case of get locator chunk zapi from
zclient, zebra precreated a down state locator and set
the chunk ownership. After this PR, this is no longer
done, and chunks are no longer automatically generated.
In this commit, we will make a test update to check the
corresponding detailed behavior.
Signed-off-by: Hiroki Shirokura <slank.dev@gmail.com>
Description:
Change is intended for fixing the following issues related to vrf route leaking:
Routes with special nexthops i.e. blackhole/sink routes when imported,
are not programmed into the FIB and corresponding nexthop is set as 'inactive',
nexthop interface as 'unknown'.
While importing/leaking routes between VRFs, in case of special nexthop(ipv4/ipv6)
once bgp announces route(s) to zebra, nexthop type is incorrectly set as
NEXTHOP_TYPE_IPV6_IFINDEX/NEXTHOP_TYPE_IFINDEX
i.e. directly connected even though we are not able to resolve through an interface.
This leads to nexthop_active_check marking nexthop !NEXTHOP_FLAG_ACTIVE.
Unable to find the active nexthop(s), route is not programmed into the FIB.
Whenever BGP leaks routes, set the correct nexthop type, so that route gets resolved
and correctly programmed into the FIB, in the imported vrf.
Co-authored-by: Kantesh Mundaragi <kmundaragi@vmware.com>
Signed-off-by: Iqra Siddiqui <imujeebsiddi@vmware.com>
Create a pid file for the router created by topotest.
By executing nsenter directly against this pid, developers
can execute commands directly from outside the unet shell.
This allows the developer to use script, tab completion, etc.,
and improves efficiency.
Signed-off-by: Hiroki Shirokura <slank.dev@gmail.com>
Refactor the bgp_auth test to create common_config code to allow
non-json based tests to reset routers and load configs in parallel.
Signed-off-by: Christian Hopps <chopps@labn.net>
- Reduce OSPF timers to 1 and 4
- Reduce BGP connect timer to 5
- Apply configs in parallel as single file
- Remove the switches as all links are p2p, perhaps this will help with
reliability?
Signed-off-by: Christian Hopps <chopps@labn.net>
Utilizes new pytest fixtures to completely factor out setup and teardown
functionality. Supply the JSON config and write your tests.
"The best topotest template yet!"
Signed-off-by: Christian Hopps <chopps@labn.net>
New generic script uses a new default node specific log dir to avoid
collisions when running in parallel.
Signed-off-by: Christian Hopps <chopps@labn.net>
- The PIM tests do not need kernel routes to help them bind joins and
sources to specific interfaces. They should do that themselves directly.
Also do not change system wide "rp_filter" sysctl away from the value
required by everyone else.
Signed-off-by: Christian Hopps <chopps@labn.net>
There were some tests where we were turning on mpls on
interface names that don't exist for certain `machines`
in the topology. Fix.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
This is to avoid breaking changes between existing deployments of
extended community for bandwidth encoding. By default FRR uses uint32
to encode bandwidth, which is not as the draft requires (IEEE floating-point).
This switch enables the required encoding per-peer.
Signed-off-by: Donatas Abraitis <donatas.abraitis@gmail.com>
This allows defining a CLI command like this:
`[no] some setting ![VALUE]`
with VALUE being optional for the "no" form, but required for the
positive form. It's just a `[...]` where the empty branch can only be
taken for commands starting with `no`.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Add the "metric" and "metric-type" options to the "redistribute"
command.
This is a small commit since the logic of setting the metric
value and type of external routes was already present due to the
implementation of the "default-information originate" command months
ago. This commit merely extends the "redistribute" command to
leverage that functionality.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Also, update the ospf6_topo2 topotest since the expected output
was wrong. With this fix, NSSA routes will be created on r2
("redistribute connected"), and NSSA routes appear in the routing
table as regular external routes.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
OSPF mixes uses of "delete" and "del_action" depending on which library
function is called. It's a bug-prone mess that needs fixing; however, for
now we fix the one obvious incorrect use in this test.
Signed-off-by: Christian Hopps <chopps@labn.net>
There is a possibility that the same line can be matched as a command in
some node and its parent node. In this case, when reading the config,
this line is always executed as a command of the child node.
For example, with the following config:
```
router ospf
network 193.168.0.0/16 area 0
!
mpls ldp
discovery hello interval 111
!
```
Line `mpls ldp` is processed as command `mpls ldp-sync` inside the
`router ospf` node. This leads to a complete loss of `mpls ldp` node
configuration.
To eliminate this issue and all possible similar issues, let's print an
explicit "exit" at the end of every node config.
This commit also changes indentation for a couple of existing exit
commands so that all existing commands are on the same level as their
corresponding node-entering commands.
Fixes#9206.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
In particular, the fixed 2 second sleep here was not long enough.
Switch to standard run_and_expect polling to make test more robust.
Signed-off-by: Christian Hopps <chopps@labn.net>
- In order to run tests in parallel the netns-based vrfs need to
have unique names primarily bc they are all tracked/looked-up in
`/run/netns` which is not network namespace nesting friendly
- use ip(8) exclusively rather than a mix of `ip` and `ifconfig`
and `vconfig`, reducing required pkg count by a couple.
Signed-off-by: Christian Hopps <chopps@labn.net>
- bugs in the support library function `verify_gr_address_family`
allowed this test to pass depending on ordering of python dictinoary
keys. Fix the bugs, fix the test.
Signed-off-by: Christian Hopps <chopps@labn.net>
Related: http://docs.frrouting.org/projects/dev-guide/en/latest/topotests.html
Directory name for a new topotest must not contain hyphen (-) characters.
To separate words, use underscores (_). For example, tests/topotests/bgp_new_example.
Signed-off-by: Donatas Abraitis <donatas.abraitis@gmail.com>
Meanwhile we don't get all MSDP features (MSDP route validation via BGP
AS Path as described in RFC 4611 Section 2), kill one of the links of
the topology to avoid intermittent test failures due to different
traffic route.
Signed-off-by: Rafael Zalamena <rzalamena@opensourcerouting.org>
Remove a 200 second sleep from bgp-evpn-overlay-index-gateway.
There does not seem to be any evidence that this is needed
and I cannot make the test fail without it.
Fixes: #9035
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
- A more general fix for the bgp listener test which requires interfaces be
configured in the kernel when the bgpd daemons are launched.
Signed-off-by: Christian Hopps <chopps@labn.net>
The test_simple_snmp.py test starts bgp, zebra and snmpd at the
same time. Then zebra configuration is read in and interface
addresses are applied. If snmp start slower than zebra
the snmp process can properly get it's ip address to bind to
if it is faster than zebra, it will fail. Ensure that the
test has addresses before we start daemons.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When running this test on a locally loaded system I am seeing the
static route as `queued` still after 1 second. Let's just blanket
increase the timeout to something longer to give a very loaded system
more time to install the route.
Output on my test system when it was loaded:
INFO topolog.r1:topogen.py:880 vtysh result:
{
"4.5.1.0/24":[
{
"prefix":"4.5.1.0/24",
"prefixLen":24,
"protocol":"static",
"vrfId":0,
"vrfName":"default",
"selected":true,
"destSelected":true,
"distance":1,
"metric":0,
"queued":true,
"table":254,
"internalStatus":8,
"internalFlags":73,
"internalNextHopNum":1,
"internalNextHopActiveNum":1,
"uptime":"00:00:00",
"nexthops":[
{
"flags":1,
"ip":"192.168.216.3",
"afi":"ipv4",
"interfaceIndex":11,
"interfaceName":"r1-eth6",
"active":true,
"weight":1
}
]
},
I suspect 10 seconds should be enough( I would hope ).
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Tests should have low enough overhead that sending
the join/prune every 5 seconds should be sufficient
also it should allow us to converge faster in case of
dropped packets.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
frrscript_load now loads a function instead of a file, so frrscript_unload
should be renamed since it does not unload a function.
Signed-off-by: Donald Lee <dlqs@gmx.com>
- Remove incorrect requirement for `service integrated-vtysh-config`
when producing a delta.
- Add `--test-reset` option which suppresses non-parseable lines from the
produced delta
- Use new features in common_config.py
Signed-off-by: Christian Hopps <chopps@labn.net>
TMUX and Screen support when running topotests inside docker. This
allows the gdb, shell and vtysh features to correctly work even when
running the tests inside docker.
Add options:
--asan-abort :: aborts the process on ASAN errors
--strace-daemons :: strace some or all daemons
Signed-off-by: Christian Hopps <chopps@labn.net>
Some BGP updates received by BGP invite local router to
install a route through itself. The system will not do it, and
the route should be considered as not valid at the earliest.
This case is detected on the zebra, and this detection prevents
from trying to install this route to the local system. However,
the nexthop tracking mechanism is called, and acts as if the route
was valid, which is not the case.
By detecting in BGP that use case, we avoid installing the invalid
routes.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Since the common CLI code calls nb_init, allow specifying some modules
to load by overriding test_yang_models.
Signed-off-by: David Lamparter <equinox@diac24.net>
Add a new topotest that features a topology with seven routers spread
across four OSPF areas:
* 1 backbone area;
* 1 regular non-backbone area (0.0.0.1);
* 1 stub area (0.0.0.2);
* 1 NSSA area (0.0.0.3).
All routers have both GR and GR helper functionality enabled in
the configuration. The test consists of restarting each router,
one at time, and checking that all forwarding planes (and LSDBs)
are kept intact during those restarts.
A successful run takes about three minutes to finish.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Using "write memory" to save the daemons' configurations before
restarting them can cause log files to stop working correctly. Add
a new "save_config" to the kill_router_daemons() function to prevent
that from happening when saving the configurations isn't necessary.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Speedup (large topo): OLD: ~6 minutes NEW: ~1 second
(when paired with generate_support_bundle.py changes)
- Collect from each node in parallel
Bug fixes:
- sub-directory test name was the same internal pytest function name
for any test, and not the actual test name.
Signed-off-by: Christian Hopps <chopps@labn.net>
On a loaded machine running FRR with ASAN I've got the following result:
INFO: waiting MSDP connection from peer 10.254.254.3 on router r1
INFO: 'router_json_cmp' polling started (interval 1 secs, maximum 30 tries)
INFO: 'router_json_cmp' succeeded after 22.53 seconds
Which is very close to the limit, so lets bump the value 4x to avoid a
test false positive.
Signed-off-by: Rafael Zalamena <rzalamena@opensourcerouting.org>
This python fixture was way too complex for what is needed.
Eliminate gratuitous options/over-engineering:
- Change from non-deterministic `wait` and `attempts` to a single
`retry_timeout` value. This is both more deterministic, as well as
what the user should actually be thinking about.
- Use a fixed 2 second pause between executing the wrapped function
rather than a bunch of arbitrary choices of 2, 3 and 4 seconds
spread all over the test code.
- Get rid of the multiple variables for determining what "Positive" and
"Negative" results are. Instead just implement what all the user code
already wants, i.e., boolean False or a str (errormsg) means
"Negative" result otherwise it's a "Positive" result.
- As part of the above the inversion logic is much more comprehensible
in the fixture code (and more correct to boot).
Signed-off-by: Christian Hopps <chopps@labn.net>
Pylint cleanup in commit 914faab594 removed a crucial function
parameter that inverted the logic of verify function calls.
Signed-off-by: Christian Hopps <chopps@labn.net>
Modify both the default and vrf ospf6 topologies to include a test
where write-multiplier is configured to a non-default value and
the ospf6 neighbors are reset then checked.
Run black on both test files.
Signed-off-by: Pat Ruddy <pat@voltanet.io>
Currently, the dynamic hostname cache is global. It is incorrect because
neighbors in different VRFs may have the same system ID and different
hostnames.
This also fixes a memory leak - when the instance is deleted, the cache
must be cleaned up and the cleanup thread must be cancelled.
Fixes#8832.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
ospf6d (and all other daemons except zebra) doesn't correctly process
`interface X vrf Y`, because it doesn't know existing VRFs at the time
of configuration file reading. Therefore it doesn't apply configuration
provided in the interface node.
Fix the problem by removing `vrf Y` part, having just an interface name
is enough.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
Add a terse option to show bgp summary to shorten output.
Do not show the following information about the BGP
instances: the number of RIB entries, the table version and the used memory.
The "terse" option can be used in combination with the "remote-as", "neighbor",
"failed" and "established" filters, and with the "wide" option as well.
Before patch:
ubuntu# show bgp summary remote-as 123456
IPv4 Unicast Summary (VRF default):
BGP router identifier X.X.X.X, local AS number XXX vrf-id 0
BGP table version 0
RIB entries 3, using 552 bytes of memory
Peers 5, using 3635 KiB of memory
Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd PfxSnt Desc
10.200.200.2 4 123456 81432 4 0 56092 0 00:00:13 572106 0 N/A
Displayed neighbors 1
Total number of neighbors 4
IPv6 Unicast Summary (VRF default):
BGP router identifier X.X.X.X, local AS number XXX vrf-id 0
BGP table version 0
RIB entries 3, using 552 bytes of memory
Peers 5, using 3635 KiB of memory
Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd PfxSnt Desc
% No matching neighbor
Total number of neighbors 5
After patch:
ubuntu# show bgp summary remote-as 123456 terse
IPv4 Unicast Summary (VRF default):
BGP router identifier X.X.X.X, local AS number XXX vrf-id 0
Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd PfxSnt Desc
10.200.200.2 4 123456 81432 4 0 56092 0 00:00:13 572106 0 N/A
Displayed neighbors 1
Total number of neighbors 4
IPv6 Unicast Summary (VRF default):
BGP router identifier X.X.X.X, local AS number XXX vrf-id 1
% No matching neighbor
Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
When filtering sessions on show bgp summary with failed, established,
neighbor and remote-as options, add a counter of displayed neighbors
in addition to the total number of neighbor :
"Displayed neighbors X"
ubuntu# show bgp summary failed remote-as external
IPv4 Unicast Summary (VRF default):
Neighbor EstdCnt DropCnt ResetTime Reason
10.200.200.2 0 0 never Waiting for NHT
172.16.29.2 0 0 never Waiting for NHT
10.22.1.2 0 0 never Waiting for NHT
Displayed neighbors 3
Total number of neighbors 5
Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
Display on which VRF/view the neighbor was not found. Useful when
selecting "vrf all".
Before patch:
No such neighbor in this view/vrf
After patch:
No such neighbor in VRF default
Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
Issue: There was an error reported by Pylint regarding "expected" keyword:
Unexpected keyword argument 'expected' in function call (unexpected-keyword-arg)
Fix: We have defined expected keyword in all topojson APIs.
Signed-off-by: Kuldeep Kashyap <kashyapk@vmware.com>
Following functionality is covered:
+--------+ BGP +--------+ BGP +--------+ +--------+
SN1 | | IPv4/v6 | | EVPN | | | |
======+ Host1 +---------+ PE1 +------+ PE2 +------+ Host2 +
| | | | | | | |
+--------+ +--------+ +--------+ +--------+
Host1 is connected to PE1 and host2 is connected to PE2
Host1 and PE1 have IPv4/v6 BGP sessions.
PE1 and PE2 gave EVPN session.
Host1 advertises IPv4/v6 prefixes to PE1.
PE1 advertises these prefixes to PE2 as EVPN type-5 routes.
Gateway IP for these EVPN type-5 routes is host1 IP.
Host1 MAC/IP is advertised by PE1 as EVPN type-2 route
Following testcases are covered:
TC_1:
Check BGP and zebra states for above topology at PE1 and PE2.
TC_2:
Stop advertising prefixes from host1. It should withdraw type-5 routes. Check
states at PE1 and PE2
Advertise the prefixes again. Check states.
TC_3:
Shut down VxLAN interface at PE1. This should withdraw type-2 routes. Check
states at PE1 and PE2.
Enable VxLAN interface again. Check states.
Signed-off-by: Ameya Dharkar <adharkar@vmware.com>
To start we use 10k static route config. This test goes along with
recent batching changes it will fail w/o them (b/c some operations w/o
batching take 100 times as long).
This test should be added to over time for other large config
items (e.g., acl, policy, etc)
Signed-off-by: Christian Hopps <chopps@labn.net>
Test uses staticd which required some C++ header protections.
Additionally, the test also runs in the ubuntu20 docker container as
grpc is supported there by the packaging system.
Signed-off-by: Christian Hopps <chopps@labn.net>
similarly to what was done for IS-IS in commit 01d43141, combine
the SRGB and SRLB commands for OSPF-SR, so that we can replace
overlapping ranges in one sweep change.
Also allow the range configuration to be stored before SR is enabled.
There is no reason why we should not - in fact that constraint meant
that we were always requesting the default label ranges regardless
of what we actually wanted to use.
Finally, update the topotests now that we do not need to refresh
the SRGB/SRLB/MSD after disabling SR. Note that the prefix-sid still
needs to be re-added.
Signed-off-by: Emanuele Di Pascale <emanuele@voltanet.io>
Modify VRF/view display in show bgp summary:
- to be more concise
- to display on which VRF/view no neighbor was found
Before patch:
ubuntu# show bgp vrf all summary
Instance default:
IPv4 Unicast Summary:
BGP router identifier XX.XX.XX.XX, local AS number XXXX vrf-id 0
(...)
IPv6 Unicast Summary:
Instance private:
IPv4 Unicast Summary:
ubuntu# show bgp vrf all ipv4 multicast summary
% No BGP neighbors found
% No BGP neighbors found
After patch:
ubuntu# show bgp vrf all summary
IPv4 Unicast Summary (VRF default):
BGP router identifier XX.XX.XX.XX, local AS number XXXX vrf-id 0
(...)
IPv6 Unicast Summary (VRF default):
(...)
IPv4 Unicast Summary (VRF private):
(...)
ubuntu# show bgp vrf all ipv4 multicast summary
% No BGP neighbors found in VRF default
% No BGP neighbors found in VRF private
Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
New OSPFv3 NSSA test:
* When a static route is redistributed to an NSSA router it should be
type 7 and should show up in OSPFv3 route database.
* Test LSA Type 7 and route removal.
Co-authored-by: Soman K.S <somanks@gmail.com>
Signed-off-by: Rafael Zalamena <rzalamena@opensourcerouting.org>
https://github.com/FRRouting/frr/pull/5865#discussion_r597670225
As this comment says. ZEBRA_FLAG_XXX should not have been used.
To communicate SRv6 Route Information. A simple Nexthop Flag would
have been sufficient for SRv6 information. And I fixed the whole
thing that way.
Signed-off-by: Hiroki Shirokura <slank.dev@gmail.com>
The "show sharp segment-routing srv6" command was a
json output command, but it did not follow the common
practice of the other commands.
It follows the review and outputs the json format by
using the json keyword. Otherwise, it produces human
readable output.
Signed-off-by: Hiroki Shirokura <slank.dev@gmail.com>
This commit fix bgpd's prefix-sid type4,5 feature which has
miss implementation from https://github.com/FRRouting/frr/pull/5653
was merged. Due to some nessesary lines are not presented.
When bgpd receives multi update message with same service-sid on
prefix-sid type-5 attribute, bgpd will crash arround path-attribute's
values object reference count.
And also, this commit add a topotest to check that feature work fine.
Signed-off-by: Hiroki Shirokura <slank.dev@gmail.com>
This commit is a part of #5853 works.
This commit add new topotest to verify SRv6-manager's functionality.
Following tests are performed on this topotest.
- check that SRv6-locator is set correctly
- check that default SRv6-function locator is set correctly
- check that SRv6-function is installed as ipv6 route correctly
Signed-off-by: Hiroki Shirokura <slank.dev@gmail.com>
This commit checks seg6local route configuration via ZAPI is
working fine.
SRv6 feature is little young feature as kernel feature so netlink
interface may be changed/updated in the future. And this ZAPI extention
is something to support new routing paradigm, so it should be checked by
topotests until srv6 feature of linux kernel will be well stable.
Signed-off-by: Hiroki Shirokura <slank.dev@gmail.com>
The backoff code assumed that yang operations always completed quickly.
It checked for > 100 YANG modeled commands happening in under 1 second
to enable batching. If 100 yang modeled commands always take longer than
1 second batching is never enabled. This is the exact opposite of what
we want to happen since batching speeds the operations up.
Here are the results for libyang2 code without and with batching.
| action | 1K rts | 2K rts | 1K rts | 2K rts | 20k rts |
| | nobatch | nobatch | batch | batch | batch |
| Add IPv4 | .881 | 1.28 | .703 | 1.04 | 8.16 |
| Add Same IPv4 | 28.7 | 113 | .590 | .860 | 6.09 |
| Rem 1/2 IPv4 | .376 | .442 | .379 | .435 | 1.44 |
| Add Same IPv4 | 28.7 | 113 | .576 | .841 | 6.02 |
| Rem All IPv4 | 17.4 | 71.8 | .559 | .813 | 5.57 |
(IPv6 numbers are basically the same as iPv4, a couple percent slower)
Clearly we need this. Please note the growth (1K to 2K) w/o batching is
non-linear and 100 times slower than batched.
Notes on code: The use of the new `nb_cli_apply_changes_clear_pending`
is to commit any pending changes (including the current one). This is
done when the code would not correctly handle a single diff that
included the current changes with possible following changes. For
example, a "no" command followed by a new value to replace it would be
merged into a change, and the code would not deal well with that. A good
example of this is BGP neighbor peer-group changing. The other use is
after entering a router level (e.g., "router bgp") where the follow-on
command handlers expect that router object to now exists. The code
eventually needs to be cleaned up to not fail in these cases, but that
is for future NB cleanup.
Signed-off-by: Christian Hopps <chopps@labn.net>
There are two possible use-cases for the `vrf_bind` function:
- bind socket to an interface in a vrf
- bind socket to a vrf device
For the former case, there's one problem - success is returned when the
interface is not found. In that case, the socket is left unbound without
throwing an error.
For the latter case, there are multiple possible problems:
- If the name is not set, then the socket is left unbound (zebra, vrrp).
- If the name is "default" and there's an interface with that name in the
default VRF, then the socket is bound to that interface.
- In most daemons, if the router is configured before the VRF is actually
created, we're trying to open and bind the socket right after the
daemon receives a VRF registration from zebra. We may not receive the
VRF-interface registration from zebra yet at that point. Therefore,
`if_lookup_by_name` fails, and the socket is left unbound.
This commit fixes all the issues and updates the function description.
Suggested-by: Pat Ruddy <pat@voltanet.io>
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
Add ability to filter session on show bgp summary by neighbor or
remote AS:
ubuntu# show bgp summary ?
neighbor Show only the specified neighbor session
remote-as Show only the specified remote AS session
ubuntu# show bgp summary neighbor ?
A.B.C.D Neighbor to display information about
WORD Neighbor on BGP configured interface
X:X::X:X Neighbor to display information about
ubuntu# show bgp summary remote-as ?
(1-4294967295) AS number
external External (eBGP) AS sessions
internal Internal (iBGP) AS sessions
This patch includes the documentation and the topotest.
Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
Test case test_verify_mroute_when_5_different_receiver_joining_same_sources_p0
is failing intermittently in master. Fixed the issue.
Signed-off-by: Kuldeep Kashyap <kashyapk@vmware.com>
1. Automated test cases to verify BGP Graceful Shutdown community functionality,
with 2 different topologies.
Signed-off-by: Kuldeep Kashyap <kashyapk@vmware.com>
We only need an instance when we have at least one area configured in a
VRF. Currently we have the following issues:
- instance for the default VRF is always created
- instance is not removed after the last area config is removed
This commit fixes both issues.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
When the redistribution is configured in non-default VRF, isisd should
redistribute routes from this VRF instead of default.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
Compile with v2.0.0 tag of `libyang2` branch of:
https://github.com/CESNET/libyang
staticd init load time of 10k routes now 6s vs ly1 time of 150s
Signed-off-by: Christian Hopps <chopps@labn.net>
Change every `-` to `_` in directory names. This is to avoid mixing _ and -.
Just for consistency and directory sorting properly.
Signed-off-by: Donatas Abraitis <donatas.abraitis@gmail.com>
BGP_MAX_PACKET_SIZE no longer represented the absolute maximum BGP
packet size as it did before, instead it was defined as 4096 bytes,
which is the maximum unless extended message capability is negotiated,
in which case the maximum goes to 65k.
That introduced at least one bug - last_reset_cause was undersized for
extended messages, and when sending an extended message > 4096 bytes
back to a peer as part of NOTIFY data would trigger a bounds check
assert.
This patch redefines the macro to restore its previous meaning,
introduces a new macro - BGP_STANDARD_MESSAGE_MAX_PACKET_SIZE - to
represent the 4096 byte size, and renames the extended size to
BGP_EXTENDED_MESSAGE_MAX_PACKET_SIZE for consistency. Code locations
that definitely should use the small size have been updated, locations
that semantically always need whatever the max is, no matter what that
is, use BGP_MAX_PACKET_SIZE.
BGP_EXTENDED_MESSAGE_MAX_PACKET_SIZE should only be used as a constant
when storing what the negotiated max size is for use at runtime and to
define BGP_MAX_PACKET_SIZE. Unless there is a future standard that
introduces a third valid size it should not be used for any other
purpose.
Signed-off-by: Quentin Young <qlyoung@nvidia.com>
There is a rare case where with prefix peers the peer is
completely absent from the json output when checking the
peer state resulting in a python key error. Check key exists
before checking the state.
Signed-off-by: Pat Ruddy <pat@voltanet.io>
parse_topology function doesn't correctly process vertex types with
spaces. Therefore the reference topology files are completely messed up,
we have values in incorrect fields - types in metrics, metrics in
parents, etc.
This commit fixes the parsing function and the reference files.
The same fix was done for isis-topo1-vrf in #8365.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
Problem Statement:
=================
In scale setup BGP sessions start flapping.
RCA:
====
In virtualized environment there are multiple places where
MTU need to be set. If there are some places were MTU is not set
properly then there is chances that BGP packets get fragmented,
in scale setup this will lead to BGP session flap.
Fix:
====
A new tcp option is provided as part of this implementation,
which can be configured per neighbor and helps to set the TCP
max segment size. User need to derive the path MTU between the BGP
neighbors and set that value as part of tcp-mss setting.
1. CLI Configuration:
[no] neighbor <A.B.C.D|X:X::X:X|WORD> tcp-mss (1-65535)
2. Running config
frr# show running-config
router bgp 100
neighbor 198.51.100.2 tcp-mss 150 => new entry
neighbor 2001:DB8::2 tcp-mss 400 => new entry
3. Show command
frr# show bgp neighbors 198.51.100.2
BGP neighbor is 198.51.100.2, remote AS 100, local AS 100, internal link
Hostname: frr
Configured tcp-mss is 150, synced tcp-mss is 138 => new display
4. Show command json output
frr# show bgp neighbors 2001:DB8::2 json
{
"2001:DB8::2":{
"remoteAs":100,
"bgpTimerKeepAliveIntervalMsecs":60000,
"bgpTcpMssConfigured":400, => new entry
"bgpTcpMssSynced":388, => new entry
Risk:
=====
Low - This is a config driven feature and it sets the max segment
size for the TCP session between BGP peers.
Tests Executed:
===============
Have done manual testing with three router topology.
1. Executed basic config and un config scenarios
2. Verified if the config is updated in running config
during config and no config operation
3. Verified the show command output in both CLI format and
JSON format.
4. Verified if TCP SYN messages carry the max segment size
in their initial packets.
5. Verified the behaviour during clear bgp session.
6. done packet capture to see if the new segment size
takes effect.
Signed-off-by: Abhinay Ramesh <rabhinay@vmware.com>
Not having scapy in the docker image leads to very obtuse failures in
the pim bsm tests (obtuse, as in, it just fails without any hint as to
why...)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
The previous method, using zassert.h and hoping nothing includes
assert.h (which, on glibc at least, just does "#undef assert" and puts
its own definition in...) was fragile - and actually broke undetected.
Just provide our own assert.h and control overriding by putting it in a
separate directory to add to the include path (or not.)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Individual tests must not depend on each other. In particular, a test
can't be sure that the previous test config is applied or cleared.
It is definitely not true when a single test is executed, for example:
`test_bgp_auth.py::test_prefix_peer_remove_passwords`.
This commit makes all tests independent of each other. It also adds a
call to check_all_peers_established at the start of "remove_passwords"
tests to make sure that we not only block new peers with an incorrect
password, but also clean the existing peers.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
