Sam Hague [Thu, 14 May 2015 11:24:36 +0000 (11:24 +0000)]
Merge "Refactor OpenVSwitchUpdateCommand to improve readability."
Flavio Fernandes [Thu, 14 May 2015 01:20:03 +0000 (01:20 +0000)]
Merge "Allow terminationPoints to not need an interfaceType"
Sam Hague [Thu, 14 May 2015 01:03:22 +0000 (01:03 +0000)]
Merge "Bug 3136: Delete output port from instruction should use order instead of index"
Sam Hague [Thu, 14 May 2015 01:02:36 +0000 (01:02 +0000)]
Merge "Bug 3147 OVSDB Bridge update external_ids"
Sam Hague [Thu, 14 May 2015 00:56:58 +0000 (20:56 -0400)]
Allow terminationPoints to not need an interfaceType
Change-Id: I3cb09cb16535fc5e23e3a5f976d0095a16208db2
Signed-off-by: Sam Hague <shague@redhat.com>
(cherry picked from commit
82fdc597e0aa3d66481432d91a18646ed1fde1b7)
Ed Warnicke [Wed, 13 May 2015 22:20:43 +0000 (15:20 -0700)]
Refactor OpenVSwitchUpdateCommand to improve readability.
Note: This is *only* extracting methods so we can keep track
of what we are doing.
Change-Id: Ia09180f131c77a6ba9b00fb73009f0c7c3fe3d5b
Signed-off-by: Ed Warnicke <hagbard@gmail.com>
Ryan Goulding [Thu, 7 May 2015 11:48:45 +0000 (07:48 -0400)]
Bug 3147 OVSDB Bridge update external_ids
Fix to properly propagate external_ids and other updates to OVS. SouthboundIT
tests are added to test this functionality for bridge external_ids.
Change-Id: I7aded71307f87c0d19165367b8336b3cb1f64f14
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>
Flavio Fernandes [Thu, 7 May 2015 11:55:08 +0000 (20:55 +0900)]
Bug 3136: Delete output port from instruction should use order instead of index
Added comments to better explain the impmlementation.
The removeOutputPortFromInstructions was implemented under the assumption that
list index and order of action are equal. This could lead to action list where
order is not rebuilt properly.
Also, if removing port is not found in action list, the implementation would
cause the removal of the flow; when in fact it should leave list as is.
Change-Id: Id42d592db52c83897fd731f3bbef0434cc83d435
Signed-off-by: Flavio Fernandes <ffernand@redhat.com>
Also-By: shc411 <jung33@gmail.com>
Ed Warnicke [Tue, 12 May 2015 22:27:11 +0000 (15:27 -0700)]
Anil noticed there was a race condition here where we might write
an ovsdb-node to the operatational store before recording the
TransactInvoker we needed for people to be able to use it to write to
ovsdb. This fixes that by imposing the order:
1) Record the TransactInvoke for the ovsdb-node (internally)
2) Write the ovsdb-node to OPER
3) Register callbacks for monitors from ovsdb
Change-Id: I55202678521065b1a8c2f018a9f229b6be263a48
Signed-off-by: Ed Warnicke <hagbard@gmail.com>
evvy [Tue, 12 May 2015 08:37:49 +0000 (14:07 +0530)]
Fixing sonar issues. Amending earlier changes
Change-Id: I5e68b6795c10ed4ff83e7da293105525e9cf4f44
Signed-off-by: evvy <dhiraviam.natarajan@gmail.com>
Ed Warnicke [Mon, 11 May 2015 18:56:09 +0000 (11:56 -0700)]
Fix race condition in initial connection.
We must register the OvsdbConnectionCreateCommand prior
to registerCallback or we will get race conditions.
*Before* any bridge information can be successfully written
to the operational datastore, the OvsdbNode *must* be written.
The txInvoker guarantees writes occur in the order they were
provided to *it*.
If we have:
txInvoker.invoke(new OvsdbNodeCreateCommand(key, null,null));
registerCallBack();
We can guarantee that, and thus guarantee we will not have race conditions.
However, if we reverse that:
registerCallBack();
txInvoker.invoke(new OvsdbNodeCreateCommand(key, null,null));
We are guarenteed to have race conditions, as the results of the original
monitor dump get written to the datastore before the OvsdbNode gets written.
Please note: ovsdb is actively broken out of the box due to this race
conditions, so this must be corrected ASAP.
Change-Id: Id71221f7665ece5ec4b2eb9a0b39465618bada01
Signed-off-by: Ed Warnicke <hagbard@gmail.com>
Anil Vishnoi [Mon, 11 May 2015 11:23:44 +0000 (16:53 +0530)]
Remove duplication call for OvsdbNode augmentation
Change-Id: I405e09ced0747369085ffb0c81f010e10e73a0d9
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
Anil Vishnoi [Fri, 8 May 2015 21:56:55 +0000 (03:26 +0530)]
Client should be added to cache before we push the
info(OvsdbNodeAugmentation) to operational data store.
With the existing implementation we were putting task
in invoker queue to dump the data to operational data
store and then registtering for the connection callback.
After that it's adding client to cache. So if invoker
writes the data to operational data store before
it adds client to cache, any code listening to
OvsdbNodeAugmentaiton, might not be able to do any
operation if that key is not there in connection
manager cache. It works if invoker loose the race
of putting data to operational data store, it fails if
it succeed.
Change-Id: Ic73c5b7bbe0f94b04bf932b67b872099f2c27b5c
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
Ryan Goulding [Thu, 7 May 2015 11:48:45 +0000 (07:48 -0400)]
SouthboundIT tests for Bridge external_ids and other_config
Tests create, read and delete functionality for bridge external_ids and
other_config. Future changesets will utilize work here to test update
of bridges.
Change-Id: Icd45d19c58e4838415cccb06a3d4f08b117d678f
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>
Sam Hague [Thu, 7 May 2015 00:29:22 +0000 (00:29 +0000)]
Merge "Southbound-IT: add test to verify dpdk enabled switch."
Ryan Goulding [Wed, 6 May 2015 16:01:51 +0000 (12:01 -0400)]
Fix SouthboundIT by disconnecting after each test
SouthboundIT requires disconnection from OVSDB after each test.
Change-Id: Ifb916f103e94d53049f684a64b87c0ed1fc2efd6
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>
Ryan Goulding [Tue, 5 May 2015 15:22:57 +0000 (11:22 -0400)]
Southbound-IT: add test to verify dpdk enabled switch.
This test will
1. Check if it is a DPDK node.
2. Test exits if its not a DPDK node. If DPDK node then,
3. Create a netdev bridge.
4. Verifies that a netdev bridge was created.
5. Creates a dpdk port.
6. Verifies that a dpdk port was added.
7. deletes bridge and disconnects node.
8. Test fails if any of the 3-7 step fails.
Change-Id: I58e10ba18ed82193782b981ed35b22ee0e14fd62
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>
Robert Varga [Tue, 5 May 2015 14:58:01 +0000 (16:58 +0200)]
Fix javadoc JDK8 compatibility
JDK8 is more picky about javadoc correctness -- fix it up so we can pass
compilation.
Change-Id: Iae446e890afe91c6ca0b670c2ea53f18793b5e99
Signed-off-by: Robert Varga <rovarga@cisco.com>
Robert Varga [Tue, 5 May 2015 14:57:26 +0000 (16:57 +0200)]
Sort reported nodes
The order of the nodes is asserted by the unit tests -- make sure this
is predictable across JDK versions.
Change-Id: I804024fe66530aa380dbd48907268fc12a4d3085
Signed-off-by: Robert Varga <rovarga@cisco.com>
Robert Varga [Wed, 6 May 2015 09:05:40 +0000 (11:05 +0200)]
Remove explicit export-package
net-virt-providers/OF13Provider requires the
org.opendaylight.ovsdb.openstack.netvirt package to be visible. Instead
of hard-coding the exported packages, allow the bundle plugin to decide
what to export.
Change-Id: Iac66eeaf70854712b9dc5917c748d519178968e5
Signed-off-by: Robert Varga <rovarga@cisco.com>
Flavio Fernandes [Wed, 6 May 2015 00:37:58 +0000 (00:37 +0000)]
Merge "User provided bridge external-ids not making to operational/OVSDB."
Sharad Mishra [Tue, 5 May 2015 07:20:50 +0000 (00:20 -0700)]
User provided bridge external-ids not making to operational/OVSDB.
When setting up a new bridge with user provided external-ids to
config datastore, these external-ids get dropped and only the iid
shows up as bridge external-id in ovsdb/oper datastore. This patch
fixes that issue.
Change-Id: I04676e166cab96323aff0ce5747a950c111da99c
Signed-off-by: Sharad Mishra <sharad.d.mishra@intel.com>
Amit Mandke [Tue, 5 May 2015 01:15:33 +0000 (18:15 -0700)]
Fixing issue with interface update.
Currently by mistake it updates the port uuid (in operation tp) using interface uuid.
Change-Id: Ia4f3dfdb676c377dab83dda11148222640e6bf78
Signed-off-by: Amit Mandke <ammandke@cisco.com>
Sam Hague [Tue, 5 May 2015 00:57:27 +0000 (00:57 +0000)]
Merge "Fix for bridge other config in mdsal."
Sharad Mishra [Mon, 4 May 2015 09:47:46 +0000 (02:47 -0700)]
Fix for bridge other config in mdsal.
bridge other config are not showing up in operational or ovsdb.
Change-Id: Id1fd2f4f57fdd3aef1a25bf08229df26f6d3c76d
Signed-off-by: Sharad Mishra <sharad.d.mishra@intel.com>
Flavio Fernandes [Fri, 1 May 2015 19:48:51 +0000 (15:48 -0400)]
AbstractServices need unique flowIds in order to avoid rules coalescing
This is a follow up from gerrit https://git.opendaylight.org/gerrit/#/c/19410/.
When assembling different flow rules for different match, it is important
to make their flowIds unique. By not doing that, openflowPlugin will treat
them as one; and that is what gets pushed to OF capable node.
Also removed cases when writeFlow() was being invoked twice for no reason.
Change-Id: I898d777e779e61da3c9f070306915c4c37f65572
Signed-off-by: Flavio Fernandes <ffernand@redhat.com>
Ed Warnicke [Fri, 1 May 2015 17:58:52 +0000 (10:58 -0700)]
Fixed issue with connection-info->{local-ip,local-port} not being stored
Use of ConnectionInfo as a key in OvsdbConnectionManager
should only have remote{Ip,port} info... so:
1) Introduced a method suppressLocalIpPort() to SouthboundMapper
to strip localIp/Port info out of a ConnectionInfo if needed
2) Insured that *all* gets to the clients map in OvsdbConnectionManager
use that method before doing a lookup
3) Insured that *all* puts to the clients map in OvsdbConnectionManager
use that method before writing.
Also in passing fixed disconnect to just get the connectionInfo from
the ovsdbNode it has, instead of doing a convoluted dance and
removed the then no longer used createConnectionInfo(ovsdbNode) from
SouthboundMapper.
Change-Id: Idb2b16d52ec23267584c69c6e474652dd25a4ab0
Signed-off-by: Ed Warnicke <hagbard@gmail.com>
Flavio Fernandes [Thu, 30 Apr 2015 20:17:58 +0000 (16:17 -0400)]
Work around bug 3005: RoutingService needs special care to keep from rule coalescing
When rules with for RoutingService only differ on match for tun_id, the existing openflow implementation does
not distinguish them as separate rules and overrides them into a single OF entry. That is very bad for L3
forwarding, when there are more than 2 subnets, or when multiple tenants use overlapping prefixes.
Until bug 3005 is fixed, we will leverage cookie to force openflow to treat the rule as unique. Fortunately
this is easily done by using the src_tun_id itself as the cookie (i.e. no cache is needed).
Also, add segmentation info in logs for when router interface is being programmed.
vagrant@devstack-control:/vagrant/puppet/scripts$ sudo ovs-ofctl -O OpenFlow13 dump-flows br-int | grep table=60
cookie=0x0, duration=561.894s, table=60, n_packets=219, n_bytes=26397, priority=0 actions=goto_table:70
cookie=0x3ea, duration=468.025s, table=60, n_packets=0, n_bytes=0, priority=2048,ip,tun_id=0x3ea,nw_dst=3.0.0.0/24 actions=set_field:fa:16:3e:d6:0c:aa->eth_src,dec_ttl,set_field:0x3eb->tun_id,goto_table:70
cookie=0x3e9, duration=470.048s, table=60, n_packets=0, n_bytes=0, priority=2048,ip,tun_id=0x3e9,nw_dst=3.0.0.0/24 actions=set_field:fa:16:3e:d6:0c:aa->eth_src,dec_ttl,set_field:0x3eb->tun_id,goto_table:70
cookie=0x3eb, duration=467.009s, table=60, n_packets=0, n_bytes=0, priority=2048,ip,tun_id=0x3eb,nw_dst=2.0.0.0/24 actions=set_field:fa:16:3e:2d:28:89->eth_src,dec_ttl,set_field:0x3ea->tun_id,goto_table:70
cookie=0x3eb, duration=469.035s, table=60, n_packets=0, n_bytes=0, priority=2048,ip,tun_id=0x3eb,nw_dst=1.0.0.0/24 actions=set_field:fa:16:3e:25:0d:1b->eth_src,dec_ttl,set_field:0x3e9->tun_id,goto_table:70
cookie=0x3e9, duration=489.275s, table=60, n_packets=0, n_bytes=0, priority=2048,ip,tun_id=0x3e9,nw_dst=2.0.0.0/24 actions=set_field:fa:16:3e:2d:28:89->eth_src,dec_ttl,set_field:0x3ea->tun_id,goto_table:70
cookie=0x3ea, duration=488.184s, table=60, n_packets=0, n_bytes=0, priority=2048,ip,tun_id=0x3ea,nw_dst=1.0.0.0/24 actions=set_field:fa:16:3e:25:0d:1b->eth_src,dec_ttl,set_field:0x3e9->tun_id,goto_table:70
Ref: https://gist.github.com/
f3ba1edbab72b26c8d97
Ref: http://www.flaviof.com/blog/work/how-to-odl-with-openstack-part2.html
Change-Id: I7eb43b141f26b09a1c5870a7d494bc9c2cebecd8
Signed-off-by: Flavio Fernandes <ffernand@redhat.com>
Sam Hague [Thu, 30 Apr 2015 16:27:07 +0000 (16:27 +0000)]
Merge "Add testGetOvsdbNodes to it"
Sam Hague [Thu, 30 Apr 2015 16:11:51 +0000 (12:11 -0400)]
Add testGetOvsdbNodes to it
Change-Id: I33d1051bad7a220d14e569642a358ed9d929acf0
Signed-off-by: Sam Hague <shague@redhat.com>
Sam Hague [Thu, 30 Apr 2015 11:57:27 +0000 (07:57 -0400)]
Add is_connected to mdsal
Change-Id: I6ddfb22afbb2b4a27887602a8e00fa15a676d217
Signed-off-by: Sam Hague <shague@redhat.com>
Thomas Bachman [Wed, 29 Apr 2015 19:57:33 +0000 (15:57 -0400)]
Fix for NPE during GBP testing.
This addresses an NPE that happens if the
controller entry "isConnected" isn't set but
is referenced.
Signed-off-by: Thomas Bachman <tbachman@yahoo.com>
Sharad Mishra [Tue, 28 Apr 2015 08:45:21 +0000 (01:45 -0700)]
Add bridge fails on older OVS versions (< v1.10)
To recreate -
Run "ovs-vsctl add-br br0" on OVS 1.9 or lower.
The above command fails with SchemaVersionMismatchException.
Change-Id: Iee3adda5252e50e17c493373f74ee26e928f524f
Signed-off-by: Sharad Mishra <sharad.d.mishra@intel.com>
Sam Hague [Tue, 28 Apr 2015 19:35:30 +0000 (19:35 +0000)]
Merge "Add OpenFlow node reference to the bridge."
Sam Hague [Tue, 28 Apr 2015 11:20:07 +0000 (11:20 +0000)]
Merge "Update TerminationPoint operational data on ovsdb port/interface update"
Anil Vishnoi [Sat, 25 Apr 2015 00:36:30 +0000 (06:06 +0530)]
Update TerminationPoint operational data on ovsdb port/interface update
Change-Id: Id6f78e233c4c165d3f7e49a173794d08510abb65
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
Sharad Mishra [Mon, 27 Apr 2015 19:00:41 +0000 (12:00 -0700)]
Fix NPE in operational Bridge remove.
To Recreate -
1. Run 'ovs-vsctl add-br br0'
2. Verify that bridge is added.
3. Run 'ovs-vsctl del-br br0'
4. The logs show NPE and br0 is not removed from operational/ovsdb.
Change-Id: I742a0661e97521047ee1c16c2b0a45e3a2a1f1cd
Signed-off-by: Sharad Mishra <sharad.d.mishra@intel.com>
Sam Hague [Mon, 27 Apr 2015 11:20:25 +0000 (07:20 -0400)]
Remove assumes from southbound-it now that deletes from operational work
Change-Id: I8493086453f09b8aac6d4825543bce2d00191ec1
Signed-off-by: Sam Hague <shague@redhat.com>
Alexis de Talhouët [Thu, 23 Apr 2015 20:24:07 +0000 (16:24 -0400)]
Add OpenFlow node reference to the bridge.
Change-Id: Id000733a51f35841c5ddbd14b57887418e30f16d
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Anil Vishnoi [Fri, 24 Apr 2015 13:08:02 +0000 (18:38 +0530)]
Fixed termination point deletion from operational data store.
Plugin was not removing termination point operational data from
operational data store once port deleted from bridge
Change-Id: I61de3e9b5dcdc4ac096422f6445562a194ba9fa1
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
Anil Vishnoi [Sat, 18 Apr 2015 21:30:48 +0000 (14:30 -0700)]
Added detailed description to external-ids and other-configs in ovsdb.yang
Detailed descriptions from OpenVswitch schema were added to external-ids
and other-configs in ovsdb.yang file.
But, the javadoc created as a result of new description is not correctly
formatted. Any help here will be appreciated.
Patch 2: Fixed whitespaces
Change-Id: I0016fd02a76829fe4aa20a6967eb21c5a379eb06
Signed-off-by: Sharad Mishra <sharad.d.mishra@intel.com>
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
Sam Hague [Fri, 24 Apr 2015 02:05:52 +0000 (02:05 +0000)]
Merge "Fix for exception thrown when no ovs_version set."
Sam Hague [Fri, 24 Apr 2015 01:33:49 +0000 (01:33 +0000)]
Merge "SouthboundIT: Test to check ovs version for node."
Cédric Ollivier [Thu, 23 Apr 2015 17:03:40 +0000 (19:03 +0200)]
Delete redundant tests in Network, Port and Subnet handlers
This removes redundant tests according to Neutron*Northbound.java.
HTTP_OK is returned now if the neutron resource can be created, updated
or deleted.
Change-Id: I621121f020123653da8ef87fef62a27ab951629e
Signed-off-by: Cédric Ollivier <ollivier.cedric@gmail.com>
Cédric Ollivier [Tue, 3 Feb 2015 09:14:48 +0000 (10:14 +0100)]
Prevent the network shared attribute from being updated
This modifies canUpdateNetwork() to check the network shared
attribute.
Change-Id: Ie875c4cc437c27f6b202e4e26dc1acd1aeff41f8
Signed-off-by: Cédric Ollivier <ollivier.cedric@gmail.com>
Sharad Mishra [Thu, 23 Apr 2015 16:52:38 +0000 (09:52 -0700)]
Fix for exception thrown when no ovs_version set.
I am running a DPDK accelerated OVS that was manually built.
It does not have its ovs_version field set and ODL throws exception.
Not sure if we support such ovs ... is it a support statement issue or bug ??
Change-Id: I778c89e241120058b380c96b2b074eab255a85bd
Signed-off-by: Sharad Mishra <sharad.d.mishra@intel.com>
Sharad Mishra [Thu, 23 Apr 2015 15:16:37 +0000 (08:16 -0700)]
SouthboundIT: Test to check ovs version for node.
Starting small ... this patch checks ovs version attribute
of ovs node.
I could have added this check to "testAddDeleteOvsdbNode', but decided
that it could lead to minor confusion.
Change-Id: Iab4bf14215a2ced55f3c05be5f5a083db190f286
Signed-off-by: Sharad Mishra <sharad.d.mishra@intel.com>
Ryan Goulding [Wed, 22 Apr 2015 15:03:58 +0000 (11:03 -0400)]
SouthboundIT for vlan, vlanmode and vlan trunks
Add integration tests to verify vlan, vlan mode and vlan trunks functionality.
Change-Id: I745ba976a244b42d7f38a9d43db8eacac77ba416
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>
Flavio Fernandes [Tue, 21 Apr 2015 19:51:11 +0000 (19:51 +0000)]
Merge "additional southbound-it for port/interface"
Ryan Goulding [Tue, 21 Apr 2015 15:43:34 +0000 (11:43 -0400)]
additional southbound-it for port/interface
Adds integration tests to verify ofPortRequest, termination point options,
interface external-ids, port external-ids, interface other-configs, and port
other-configs.
Change-Id: Ibf76a20a5844885e404b53d4a9c163452ca4084a
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>
Sam Hague [Tue, 21 Apr 2015 14:54:40 +0000 (10:54 -0400)]
Update poms for Maven Central Requirements
Change-Id: I73fa6324ea917594e5cd110acd8f7b68102b3ca1
Signed-off-by: Sam Hague <shague@redhat.com>
Ryan Goulding [Mon, 20 Apr 2015 21:14:57 +0000 (17:14 -0400)]
ofPort test in SouthboundIT
Adds an integration test to verify ofPort functionality. A helper method is
provided to create a new TerminationPoint for testing.
Change-Id: I4588b9c95dff067007ebec47ea268c0f4e1f1bdf
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>
Anil Vishnoi [Mon, 20 Apr 2015 15:28:20 +0000 (20:58 +0530)]
Fix ovsdb node disconnect - Removing ovsdb node from config data store
is not disconnecting the node.
Change-Id: I15012381814b80b887f4fb0cb5b62dbcbe9eb458
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
Ben Eze [Sun, 19 Apr 2015 23:47:24 +0000 (19:47 -0400)]
Add Unit tests for openstack/netvirt/providers/openflow13/OF13Provider.java
Change-Id: I7b92343cfb7653e44755bb54afcc003b1adebcb9
Signed-off-by: Ben Eze <beze@inocybe.ca>
Sam Hague [Sun, 19 Apr 2015 19:30:00 +0000 (15:30 -0400)]
Fix broken SouthboundIT test
The mdsalUtils field needs to be static becuase each @Test method is a totally new allocation and the value is reaallocated for each test.
Also added some extra debug logs.
Patchset 2: checkstyle
Change-Id: Ic75b5bc25cdf62b1c3daa4ee2fdea487804fb0a9
Signed-off-by: Sam Hague <shague@redhat.com>
Sam Hague [Sun, 19 Apr 2015 17:55:00 +0000 (17:55 +0000)]
Merge "TerminationPointUpdate command to update termination point on ovsdb node"
Prateek Garg [Fri, 17 Apr 2015 19:55:31 +0000 (12:55 -0700)]
TerminationPointUpdate command to update termination point on ovsdb node
Change-Id: I17c4bd890a120b4fbf74ff5288f97bf00fd224df
Signed-off-by: Prateek Garg <prategar@cisco.com>
Sam Hague [Sat, 18 Apr 2015 01:22:37 +0000 (01:22 +0000)]
Merge "Remove .../distribution/opendaylight-karaf/. It is now .../karaf"
Flavio Fernandes [Fri, 17 Apr 2015 21:20:46 +0000 (21:20 +0000)]
Merge "Add JUnit testing for L2RewriteService class."
Flavio Fernandes [Fri, 17 Apr 2015 21:20:35 +0000 (21:20 +0000)]
Merge "Add JUnit testing for L2FowardingService class."
Flavio Fernandes [Fri, 17 Apr 2015 21:20:28 +0000 (21:20 +0000)]
Merge "Add JUnit testing for LoadBalancerService class."
Alexis de Talhouët [Fri, 17 Apr 2015 14:40:10 +0000 (10:40 -0400)]
Add JUnit testing for IngressAclService class.
Change-Id: I44c088d640413d978cbbec9fa67c8a5d7d0aea8b
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Alexis de Talhouët [Fri, 17 Apr 2015 18:46:27 +0000 (14:46 -0400)]
Add JUnit testing for LoadBalancerService class.
Change-Id: I18fb554389d4baa05e0f58a538243cdaa456cc5a
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Alexis de Talhouët [Fri, 17 Apr 2015 17:18:09 +0000 (13:18 -0400)]
Add JUnit testing for L2FowardingService class.
Change-Id: I1d7b565f1cfa81b78423827d277e8cd5929dd51a
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Alexis de Talhouët [Fri, 17 Apr 2015 19:08:38 +0000 (15:08 -0400)]
Add JUnit testing for L2RewriteService class.
Change-Id: I902ce652c3edb0857cca697ead04a9ffd475db67
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Flavio Fernandes [Fri, 17 Apr 2015 20:44:46 +0000 (16:44 -0400)]
Remove .../distribution/opendaylight-karaf/. It is now .../karaf
Change-Id: I7287ac40e9a2cfc25e3551662cde4a32bf0b88e8
Signed-off-by: Flavio Fernandes <ffernand@redhat.com>
Sam Hague [Fri, 17 Apr 2015 20:20:25 +0000 (20:20 +0000)]
Merge "Move ovsdb distribution to use karaf-parent"
Sam Hague [Fri, 17 Apr 2015 19:31:27 +0000 (15:31 -0400)]
Move ovsdb distribution to use karaf-parent
Change-Id: Ief6939eb8f1cc15f1042ecc1fd7e549c60de57cc
Signed-off-by: Sam Hague <shague@redhat.com>
Flavio Fernandes [Fri, 17 Apr 2015 18:43:07 +0000 (18:43 +0000)]
Merge "Add JUnit testing for OutboundNatService class."
Flavio Fernandes [Fri, 17 Apr 2015 18:42:58 +0000 (18:42 +0000)]
Merge "Add JUnit testing for L3FowardingService class."
Flavio Fernandes [Fri, 17 Apr 2015 18:42:17 +0000 (18:42 +0000)]
Merge "Add JUnit testing for EgressAclService class."
Flavio Fernandes [Fri, 17 Apr 2015 18:40:36 +0000 (18:40 +0000)]
Merge "Add JUnit testing for RoutingService class."
Flavio Fernandes [Fri, 17 Apr 2015 18:40:11 +0000 (18:40 +0000)]
Merge "Fix for TerminationPointCreateCommand to insert a termination point, if not present already in operational store"
Flavio Fernandes [Fri, 17 Apr 2015 18:38:47 +0000 (18:38 +0000)]
Merge "Workaround for busted controller opendaylight-karaf-empty"
Alexis de Talhouët [Fri, 17 Apr 2015 17:50:51 +0000 (13:50 -0400)]
Add JUnit testing for L3FowardingService class.
Change-Id: Icb067fabb217511a5aca155f668b492d0d823ae5
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Alexis de Talhouët [Fri, 17 Apr 2015 17:44:02 +0000 (13:44 -0400)]
Add JUnit testing for OutboundNatService class.
Change-Id: Ic3da229c61af4a412ab9031d735c1f5a5fb1cc2f
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Alexis de Talhouët [Fri, 17 Apr 2015 17:32:10 +0000 (13:32 -0400)]
Add JUnit testing for RoutingService class.
Change-Id: Ie40d1009a840f646260c676b5338088d926b73bc
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Sam Hague [Fri, 17 Apr 2015 16:33:42 +0000 (12:33 -0400)]
Workaround for busted controller opendaylight-karaf-empty
Somehow the controller opendaylight-karaf-empty is busted with all the recent karaf 3.0.3 migration. Luckily we will a local distribution that doesn't seem to have the same issues.
Change-Id: I8f90981ba684dd68c17cadf6513ef4046a5b9cfb
Signed-off-by: Sam Hague <shague@redhat.com>
Prateek Garg [Fri, 17 Apr 2015 14:06:48 +0000 (07:06 -0700)]
Fix for TerminationPointCreateCommand to insert a termination point, if not present already in operational store
Change-Id: I8ed209033986d8338cc65589d8f497a152dd2011
Signed-off-by: Prateek Garg <prategar@cisco.com>
Alexis de Talhouët [Fri, 17 Apr 2015 13:57:17 +0000 (09:57 -0400)]
Add JUnit testing for EgressAclService class.
Change-Id: I9e2e0e6174a381c4a49bb667c33ec1a593dc597a
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Alexis de Talhouët [Fri, 17 Apr 2015 14:21:51 +0000 (10:21 -0400)]
Add JUnit testing for InboundNatService class.
Change-Id: Ia2a80aa73c7cabcfe91acb1fa1ab3f6718f08ca0
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Flavio Fernandes [Fri, 17 Apr 2015 13:48:38 +0000 (13:48 +0000)]
Merge changes Ifb9ee4a4,Id0a7512b
* changes:
Correct spelling of getBrideTermiationPoint()
Add bridge integration test
Sam Hague [Fri, 17 Apr 2015 12:27:11 +0000 (08:27 -0400)]
Correct spelling of getBrideTermiationPoint()
Change-Id: Ifb9ee4a4793a57a854f685cd0afce79d1ca96d2f
Signed-off-by: Sam Hague <shague@redhat.com>
Sam Hague [Fri, 17 Apr 2015 12:22:55 +0000 (12:22 +0000)]
Merge "Add Unit Tests for openstack.net-virt-providers.openflow13 - FlowCapableNodeDataChangeListener.java and MdsalConsumerImpl.java"
Sam Hague [Fri, 17 Apr 2015 12:21:26 +0000 (12:21 +0000)]
Merge "Add JUnit testing for ArpResponderService class."
Sam Hague [Fri, 17 Apr 2015 11:36:16 +0000 (07:36 -0400)]
Add bridge integration test
Change-Id: Id0a7512bc4fb08c6bec4533eb35447cf4ba57be1
Signed-off-by: Sam Hague <shague@redhat.com>
Ben Eze [Fri, 17 Apr 2015 02:05:45 +0000 (22:05 -0400)]
Add Unit Tests for openstack.net-virt-providers.openflow13 - FlowCapableNodeDataChangeListener.java and MdsalConsumerImpl.java
Change-Id: Icc614719be95e5b945a2bb686d509712a956c703
Signed-off-by: Ben Eze <beze@inocybe.ca>
Flavio Fernandes [Thu, 16 Apr 2015 23:42:59 +0000 (23:42 +0000)]
Merge "Add OpenFlow ID to ovsdb-bridge-attributes."
Flavio Fernandes [Thu, 16 Apr 2015 23:42:10 +0000 (23:42 +0000)]
Merge "BUG 3005: WORKAROUND issue where match cannot differ only on extensions portion"
Flavio Fernandes [Thu, 16 Apr 2015 23:41:45 +0000 (23:41 +0000)]
Merge "Add JUnit testing for ClassifierService class."
Flavio Fernandes [Thu, 16 Apr 2015 23:40:53 +0000 (23:40 +0000)]
Merge "Update gogo versions to account for new karaf 3.0.3 constraints"
Sam Hague [Thu, 16 Apr 2015 13:46:03 +0000 (09:46 -0400)]
Update gogo versions to account for new karaf 3.0.3 constraints
Change-Id: Id062a95990a903873ab4b1f06408889f786f50fb
Signed-off-by: Sam Hague <shague@redhat.com>
Alexis de Talhouët [Thu, 16 Apr 2015 19:11:05 +0000 (15:11 -0400)]
Add JUnit testing for ClassifierService class.
Change-Id: I4f00ae666cc28ad3d438f879da4b5780c2b1ab7c
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Flavio Fernandes [Thu, 16 Apr 2015 17:40:52 +0000 (13:40 -0400)]
BUG 3005: WORKAROUND issue where match cannot differ only on extensions portion
The details on the original issue were discovered by Anil Vishnoi:
BUG-2757 : Openflowplugin manager doesn't use match extensions while comparing two flows
Ovsdb installs two flows which is different from each other just by the match extension (reg0) (actions are not used in comparison).
cookie=0x0, duration=255.353s, table=110, n_packets=8, n_bytes=648, priority=16384,reg0=0x1,tun_id=0x1,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00 actions=output:1,output:2
cookie=0x0, duration=256.115s, table=110, n_packets=9, n_bytes=1464, priority=16384,reg0=0x2,tun_id=0x1,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00 actions=output:1
So statistics manager assign the same id internally to both the flows and that's causing the repeated error message about duplicate id.
Adding match extension in flow comparison is not trivial and we cannot implement it in this release.
Change-Id: I9d9986628fcb4996a7b7aac7af327df59201904a
Also-By: Anil Vishnoi <vishnoianil@gmail.com>
Signed-off-by: Flavio Fernandes <ffernand@redhat.com>
Alexis de Talhouët [Thu, 16 Apr 2015 15:52:27 +0000 (11:52 -0400)]
Add JUnit testing for ArpResponderService class.
Change-Id: Id4509b10ca30250c83ce588ed41016b920fa0afd
Signed-off-by: Alexis de Talhouët <adetalhouet@inocybe.com>
Sam Hague [Thu, 16 Apr 2015 01:32:26 +0000 (21:32 -0400)]
Cleanup sb it
- Removed duplicated tests
- Added options to modify logging
- Added options to add extra features to load
Change-Id: I52b68e5e35b5b7b3a52da89941410d344f654ba5
Signed-off-by: Sam Hague <shague@redhat.com>
Sam Hague [Thu, 16 Apr 2015 01:11:31 +0000 (21:11 -0400)]
Cleanup some logging in sb.
Change-Id: I0dfd06163bce69e863cec2804662350cc4c44b42
Signed-off-by: Sam Hague <shague@redhat.com>
Sam Hague [Wed, 15 Apr 2015 22:36:29 +0000 (18:36 -0400)]
Fix for not finding client(key)
The connectioninfo brought in using the connectioninfo as the key for the clients. The application using the sb only have access to the remote ip and port so the key didn't match if it was using the remote ip:port and local ip:port.
Change-Id: If61faac9bb2be571c08faa95f3beb5190499a54b
Signed-off-by: Sam Hague <shague@redhat.com>
Sam Hague [Wed, 15 Apr 2015 20:25:30 +0000 (20:25 +0000)]
Merge "Add IT for getting Open_vSwitch other_config:local_ip"
Sam Hague [Wed, 15 Apr 2015 20:25:23 +0000 (20:25 +0000)]
Merge "Add constant class for shared strings in SouthboundIT"