Vishal Thapar [Tue, 8 Nov 2016 18:34:46 +0000 (00:04 +0530)]
Add .tox to .gitignore
Running tox in docs folder generates .tox as output directory.
Change-Id: I212732620540b2f862f9249324a61d1bf84bc490
Signed-off-by: Vishal Thapar <vishal.thapar@ericsson.com>
Faseela K [Tue, 8 Nov 2016 12:57:23 +0000 (12:57 +0000)]
Merge "BUG 6726 : Loss of traffic during ODL Cluster reboot"
karthik.prasad [Thu, 3 Nov 2016 13:19:29 +0000 (18:49 +0530)]
BUG 6726 : Loss of traffic during ODL Cluster reboot
Since all ARP Traffics are punted to controller, when controller reboots
the packets are lost. The changes fixes traffic loss during cluster
reboot by adding ARP flows on DPN.
The changes includes addition of all NICIRA ARP Extension Actions to
ActionType.java
PS1: Removed Tabs
PS2: Incorporation of review comments
PS3: Incorporation of review comments
PS5: Removed Deprecated method and also removed whitespace.
PS6: Removal of whitespace
PS7: Incorporation of review comments
Change-Id: Ida59553187ba20b10bbc4e08d20e491d5dd72200
Signed-off-by: karthik.prasad <karthik.p@altencalsoftlabs.com>
Michael Vorburger [Tue, 8 Nov 2016 11:09:42 +0000 (12:09 +0100)]
Add missing odl-infrautils-inject to features.xml
This is required as Genius is starting to use adopt
https://wiki.opendaylight.org/view/BestPractices/DI_Guidelines with
AbstractLifecycle e.g. in https://git.opendaylight.org/gerrit/#/c/47426/
and https://git.opendaylight.org/gerrit/#/c/48036/
Change-Id: I67f18b2c867c6db614cd576ca93fea783510d3ba
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Vishal Thapar [Mon, 7 Nov 2016 16:23:04 +0000 (21:53 +0530)]
Add Docs for Genius
This patch just initializes sphinx for Genius
Refer: https://git.opendaylight.org/gerrit/#/c/39142/
Change-Id: Id851cd18a9f4d6c28f0995e7d7956b2c9e75f0d3
Signed-off-by: Vishal Thapar <vishal.thapar@ericsson.com>
Faseela K [Mon, 7 Nov 2016 15:33:58 +0000 (15:33 +0000)]
Merge "Prepare for enabling checkstyle fail on error"
Michael Vorburger [Fri, 4 Nov 2016 19:00:20 +0000 (20:00 +0100)]
alarmmanager Checkstyle clean-up (incl. exception handling fixes)
Change-Id: I10a09756e4d89f4f9e573cfda82dee4456f32991
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Sam Hague [Fri, 4 Nov 2016 18:03:58 +0000 (14:03 -0400)]
Prepare for enabling checkstyle fail on error
This patch adds the checkstyle plugin config to
fail on checkstyle errors. Uncomment the checkstyle.violationSeverity
line when ready to enable.
If just adding one bundle at a time rather than all the bundles using
the commons/parents, then copy the same plugin section to the bundle
poms and uncomment the severity. As all bundles are completed then
remove the bundle sections and just use the one in the commons/pom.xml.
Change-Id: I68919f619d84676c5c22a949283f9788dbd4da8e
Signed-off-by: Sam Hague <shague@redhat.com>
Faseela K [Thu, 3 Nov 2016 09:04:38 +0000 (09:04 +0000)]
Merge "Add info to log message with ARP response details on transmit"
Alon Kochba [Tue, 1 Nov 2016 18:31:06 +0000 (20:31 +0200)]
Add info to log message with ARP response details on transmit
Change-Id: I1b09b27d7dea657da9085f8076ce4c2009f68e30
Signed-off-by: Alon Kochba <alonko@hpe.com>
Michael Vorburger [Wed, 2 Nov 2016 10:18:28 +0000 (11:18 +0100)]
Intro. @Inject @Singleton etc. in some helper services, for testability
This should not have any impact to real runtime code, and is hopefully
low risk. In particular, this alone does *NOT* yet actually exploit
these annotations for auto. generating blueprint XML (because I'm not
activating the blueprint-maven-plugin here, yet).
But adding these is useful already, because it allows these helper
services to be easily used in dependency injection via Guice in
Component Tests.
These annotation can however, later also be used by the
blueprint-maven-plugin - once the respective projects fully switch.
This change also includes a bunch of otherwise completely unrelated
automated formatting and code clean-up which happened automatically as I
was touching these files.
Change-Id: If1c3fe6d03f089bbeee0453742fe5cea8f4f098d
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Faseela K [Wed, 2 Nov 2016 07:21:51 +0000 (07:21 +0000)]
Merge "BUG 6626 Packet IN handler thread in deadlock after high ARP rate"
Faseela K [Wed, 2 Nov 2016 07:13:43 +0000 (07:13 +0000)]
Merge "Bug 6836 - No access to external network"
Tali [Tue, 1 Nov 2016 15:18:04 +0000 (17:18 +0200)]
Bug 6836 - No access to external network
1) Additional race fix
2) Move future creation to be per worker
Change-Id: I02fe765021489c71a9c3549d2d41767c65578a50
Signed-off-by: Tali <tali.ben-meir@hpe.com>
Sam Hague [Tue, 1 Nov 2016 20:52:29 +0000 (20:52 +0000)]
Merge "target-ide/ on .gitignore"
Faseela K [Tue, 1 Nov 2016 09:17:08 +0000 (09:17 +0000)]
Merge "Migrate Resource Manager service to BluePrint"
Michael Vorburger [Mon, 31 Oct 2016 17:37:52 +0000 (18:37 +0100)]
target-ide/ on .gitignore
Change-Id: I5006d521cc46191a3a88ec162e970ced453d424a
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Tali [Thu, 27 Oct 2016 11:05:18 +0000 (13:05 +0200)]
Bug 6836 - No access to external network
Upon first neutron network VLAN provider creation with a defined physnet
both VLAN trunk VLAN member are created for each compute node in the ietf-interface config DS.
As a result ietf-interface state for the VLAN member is attampted to be created simultaneously by
1) InterfaceConfigListener - triggered as a result of VLAN trunk creation in the config DS
2) VlanMemberConfigListener - triggered as a result of VLAN member creation in the config DS
Due to race condition in ietf-interface state for the VLAN memeber creation,
OptimisticLockFailedException can be sporadically get thrown and ietf-interface state
will not be added to the oper DS.
Suggested solution:
When renderers triggered by InterfaceConfigListener on the VLAN trunk call OvsInterfaceConfig*Helpers
to update VLAN member state, these jobs will be queued based on the trunk name to the DjC to prevent
concurrent write of VLAN member ietf-interface state with VlanMemberConfigListener
Change-Id: I0a89e73c6ff0acb9bdc369ab73e16df43e2f727b
Signed-off-by: Tali <tali.ben-meir@hpe.com>
David Suarez [Mon, 24 Oct 2016 08:26:37 +0000 (10:26 +0200)]
Migrate Resource Manager service to BluePrint
Migration of the resourcemanager-impl module to use BluePrint
framework.
Change-Id: I6b6ff7fa70fa4e1ba2831bcee74428658a4bd0d4
Signed-off-by: David Suarez <david.suarez.fuentes@ericsson.com>
Vivek Srivastava [Fri, 28 Oct 2016 09:09:34 +0000 (09:09 +0000)]
Merge "Bug 7048 - Update to OF port does not change 220 flow"
Faseela K [Fri, 28 Oct 2016 09:00:39 +0000 (09:00 +0000)]
Merge "Added postman collections for id-manager"
Faseela K [Thu, 27 Oct 2016 13:41:09 +0000 (19:11 +0530)]
Bug 7048 - Update to OF port does not change 220 flow
Description : Wrong instance identifier was being used for unbind service.
Also, DJC synchronization key for unbind service should be parent interface, rather
than interface-name, else this will cause race conditions while deleting 220 flow.
Change-Id: I632791ed131cab96e1967542c1fa7e18fa92c519
Signed-off-by: Faseela K <faseela.k@ericsson.com>
N Edwin Anthony [Fri, 28 Oct 2016 06:01:35 +0000 (11:31 +0530)]
Added postman collections for id-manager
Change-Id: I0ac1ca0c739aa436ffc2886308e6be220c327666
Signed-off-by: N Edwin Anthony <n.edwin.anthony@ericsson.com>
shlomi [Wed, 26 Oct 2016 15:50:26 +0000 (17:50 +0200)]
BUG 6626 Packet IN handler thread in deadlock after high ARP rate
use offerNotification instead of putNotification. this will reject messages over the queue limit
Change-Id: Ia43ce75a8a6e7e3b1ed9e90704e03656e7cfdc11
Signed-off-by: shlomi <shlomi.alfasi@hpe.com>
Vivek Srivastava [Thu, 27 Oct 2016 08:34:18 +0000 (08:34 +0000)]
Merge "Enhancing service binding logic to support more services"
Faseela K [Mon, 24 Oct 2016 12:27:15 +0000 (17:57 +0530)]
Enhancing service binding logic to support more services
Description : Currently Genius supports only a maximum of 7 services per interface.
Enhancing this logic to support maximum 16 services per interface by making use of one
more bit from metadata field.
Change-Id: I711c03529b78924f2089300b2a52e9daa7aa5ed8
Signed-off-by: Faseela K <faseela.k@ericsson.com>
Vivek Srivastava [Thu, 27 Oct 2016 05:17:12 +0000 (05:17 +0000)]
Merge "SFC table addition; Offset the other SFC tables."
Miguel Duarte [Wed, 26 Oct 2016 16:06:14 +0000 (18:06 +0200)]
SFC table addition; Offset the other SFC tables.
Added one missing SFC table. All other SFC tables had to be offset by 1.
(now they are using indexes 82 to 87, inclusive)
Change-Id: I691ec349c47b96db182b88bf6df51110bd72504e
Signed-off-by: Miguel Duarte <miguel.duarte.de.mora.barroso@ericsson.com>
Edwin Anthony [Wed, 26 Oct 2016 09:48:28 +0000 (15:18 +0530)]
Adding resourcemanager postman collection
Change-Id: I6336049f498b1b6bcf32fdb55900ab63f0f37fa2
Signed-off-by: Edwin Anthony <n.edwin.anthony@ericsson.com>
Viji J [Wed, 26 Oct 2016 07:11:48 +0000 (12:41 +0530)]
Fix for fcaps application module config push error
Change-Id: I9f13377d332502317563bf46c8477c2fda53f238
Signed-off-by: Viji J <viji.j@ericsson.com>
Vivek Srivastava [Sun, 23 Oct 2016 15:06:36 +0000 (15:06 +0000)]
Merge "Bug 6765: Overriding in_port in table0 with Zero value"
Sam Hague [Sat, 22 Oct 2016 20:56:51 +0000 (20:56 +0000)]
Merge "Fixes for IT base"
Sam Hague [Sat, 22 Oct 2016 20:56:22 +0000 (20:56 +0000)]
Merge "Add pom for commons"
Sam Hague [Sat, 22 Oct 2016 20:55:41 +0000 (20:55 +0000)]
Merge "Bug 6278: Switch to use odlparent's karaf-parent"
xcheara [Thu, 20 Oct 2016 07:13:46 +0000 (12:43 +0530)]
Bug 6765: Overriding in_port in table0 with Zero value
Description :(Review Re-raised) OVS split-horizon logic will drop pkts
by default if value of in_port and out_port is same. But, if we have two
vlan-trunk member VM for which trunk interface is same, the connectivity
between these two VM cannot be established as OVS drops such packets.
So, it is required to override the value of in_port so that split_horizon
check is bypassed. The in_port value will be overrriden using load(ex:
load:0->NXM_OF_IN_PORT[]).
Change-Id: I3d7c457bd030b9d386037c4b6002284a83802b09
Signed-off-by: xcheara <chetan.arakere@altencalsoftlabs.com>
Sam Hague [Sat, 22 Oct 2016 15:33:03 +0000 (15:33 +0000)]
Merge "IdManager Performance Improvements"
Sam Hague [Thu, 20 Oct 2016 17:49:05 +0000 (13:49 -0400)]
Fixes for IT base
Change-Id: Idce92e17acabc948e599e992a0e669ad41198341
Signed-off-by: Sam Hague <shague@redhat.com>
Sam Hague [Thu, 20 Oct 2016 17:22:15 +0000 (13:22 -0400)]
Add pom for commons
Change-Id: I3faa6f4c36efe209b30676d96c3d56af1f615f29
Signed-off-by: Sam Hague <shague@redhat.com>
unknown [Thu, 20 Oct 2016 13:01:29 +0000 (18:31 +0530)]
Enhancements to improve DJC transaction retry mechanisms
Description : Adding linear time delay for each retries in case of job
execution failure in datastorejobcordinator
Change-Id: Icd2bd50b3dd80236f32c5be0fd371ce508dbbbf3
Signed-off-by: Faseela K <faseela.k@ericsson.com>
Kency Kurian [Thu, 29 Sep 2016 12:30:06 +0000 (18:00 +0530)]
IdManager Performance Improvements
In order to avoid the wait time to write to the DS during allocation of id
each local id pool(pool per cluster) can be stored in memory using
concurrent hash map. This can be later synced with the DS using
DatastoreJobCoordinator. In order to achieve this the following has been
done:
- Created 5 new jobs to update the local pool info DS.
* LocalPoolCreateJob - Writes the entire local pool content from the
cache to the DS.
* IdHolderSyncJob - Updates either of the idHolder(Available/Released)
* CleanUpJob - Used to return back excessive ids if any to the parent
pool. This will be scheduled after IDLE_TIME whenever an id is
released since that is the only time there are chances of excessive
ids to be present in the local pool.
* UpdateIdEntryJob - To add/delete the idEntry in the parentPool.
* LocalPoolDeleteJob - Deletes the local pool from DS.
- New classes to hold local id pool objects.
* IdHolder - An interface which provides methods to allocateId, addId
etc. which is implemented by AvailableIdHolder and ReleaseIdHolder.
* IdLocalPool - Holds the local pool name and IdHolders for each pool.
- JUnits test cases have also been updated.
- Added a CLI for printing the local cache values.
Change-Id: Ic39cae3b4c01ce44d1624187d2bf3099e707611e
Signed-off-by: Kency Kurian <kency.kurian@ericsson.com>
Ryan Goulding [Wed, 19 Oct 2016 17:26:14 +0000 (13:26 -0400)]
Bug 6278: Switch to use odlparent's karaf-parent
Change-Id: Iad5fc802e638963d547e535eace90f07ce0f8519
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>
Faseela K [Wed, 19 Oct 2016 08:34:19 +0000 (08:34 +0000)]
Merge "Upstreaming ITM cache impl and monitoring bug fix"
Sam Hague [Tue, 18 Oct 2016 20:53:21 +0000 (16:53 -0400)]
Add alivenessmonitor rpc
Change-Id: Icc8414bf23dc28d797ab7038629d056754cf0989
Signed-off-by: Sam Hague <shague@redhat.com>
Hema Gopalkrishnan [Fri, 7 Oct 2016 05:57:46 +0000 (11:27 +0530)]
Upstreaming ITM cache impl and monitoring bug fix
Removed the Add and Remove External End Point CLI from this review
Fixed caches related failures in JUnit Tests
Retaining the oper-state along with tunnel-state in tunnels_state
Change-Id: Idb783c185ce349309c8e089013302f3dc76b2d97
Signed-off-by: Hema Gopalkrishnan <hema.gopalkrishnan@ericsson.com>
David Suarez [Sat, 15 Oct 2016 16:26:47 +0000 (16:26 +0000)]
Merge "Adds abstract base for Jobs that must be validated"
Akash [Thu, 13 Oct 2016 11:07:29 +0000 (16:37 +0530)]
Bug 6791 Introduction of Async & Clustere listner
Two new listners are introducled which will be responsible for listning to
HA changes. Clustered Listner and normal Listner.
If its HA childen node or normal node then the particular
add/remove/update node will be executed accordingly.
Change-Id: Ie28137644e4fd91790fb5af697c16dcb5aec216a
Signed-off-by: Akash <a.k.sahu@ericsson.com>
Faseela K [Thu, 13 Oct 2016 05:25:21 +0000 (05:25 +0000)]
Merge "blueprint for alivenessmonitor"
Sam Hague [Fri, 7 Oct 2016 17:14:02 +0000 (13:14 -0400)]
blueprint for alivenessmonitor
Change-Id: I7a3c25ace7b8b26bbea0dc84983336c0531d8403
Signed-off-by: Sam Hague <shague@redhat.com>
Chintan Apte [Wed, 12 Oct 2016 10:36:11 +0000 (16:06 +0530)]
ODL Bug: 6095,
bundle:diag failing for ITM bundle.
UT:-
RemoveExternalEndpoint is pointing to a vpnservice package which is
causing the issue,
Started the Karaf and checked the bundle status and diag. coming up jjst
fine.
Change-Id: I76613223e5819154e6f505e42ba435500a8f5a28
Signed-off-by: Chintan Apte <chintan.apte@ericsson.com>
David Suarez [Fri, 7 Oct 2016 09:35:52 +0000 (11:35 +0200)]
Checkstyle and formatting.
Change-Id: I27c12fa9bae2333431acee6c4f7acbcc03550f2c
Signed-off-by: David Suarez <david.suarez.fuentes@ericsson.com>
Diego Granados [Wed, 5 Oct 2016 14:34:33 +0000 (16:34 +0200)]
SFC table addition + relocation
Added one missing SFC table. All SFC tables had to be rellocated
(now they are using indexes 82 to 86, both included) in order
to fit.
Change-Id: I0994994c89f751c533d16c56abe1280af7300098
Signed-off-by: Diego Granados <diego.jesus.granados.lopez@ericsson.com>
Sathish Kumar Tippeshappa [Fri, 2 Sep 2016 05:38:39 +0000 (11:08 +0530)]
Default transport zone creation during the bootup
CSC to create a default transport zone at bootup
-Any switch that connects via OVSDB with no transport zone
configured will get added into the default transport zone
-This transport zone will have default monitoring and tunnel type
settings as per default values kept in the YANG file
Testcases:-
1.Switch is connected with TZNAME="" and that node info is added to
defaultTransport zone
2.Once ODL controller is booted default transportzone is created
Change-Id: I3e510238e097e3a2eb3ab36d871477c46f01509a
Signed-off-by: Sathish Kumar Tippeshappa <sathishkumar.bt@altencalsoftlabs.com>
Faseela K [Thu, 6 Oct 2016 09:26:11 +0000 (09:26 +0000)]
Merge "BUG 6726 : Loss of traffic during ODL Cluster reboot"
Faseela K [Thu, 6 Oct 2016 04:50:28 +0000 (04:50 +0000)]
Merge "Make local variables creation and assignment in a single statement. Some other minor formatting (removing commented code, etc.)"
karthik.prasad [Mon, 26 Sep 2016 14:09:51 +0000 (19:39 +0530)]
BUG 6726 : Loss of traffic during ODL Cluster reboot
Since all ARP Traffics are punted to controller, when controller reboots
the packets are lost. The changes fixes traffic loss during cluster
reboot by adding ARP flows on DPN.
PS10: Added actionKey as an optional parameter for
getEgressActionForInterface
PS:11: Removed white spaces
PS12: Added description for RPC Yang field
Change-Id: Id75c6bc4f68e88b56a7b5b5722866161a8590dec
Signed-off-by: karthik.prasad <karthik.p@altencalsoftlabs.com>
Faseela K [Thu, 6 Oct 2016 04:48:33 +0000 (04:48 +0000)]
Merge "Arp cache feature changes"
David Suarez [Wed, 5 Oct 2016 20:37:43 +0000 (20:37 +0000)]
Merge "Refactoring: MoreObjects2 is now in yangtools.utils"
Miguel Perez [Tue, 4 Oct 2016 16:23:50 +0000 (18:23 +0200)]
Adds abstract base for Jobs that must be validated
+ Includes some state-checking stateful methods
+ Inheriting from Observable so that classes that enqueued the
job can be informed about job changes.
Change-Id: I249b50b6b719e13a5e3e70988c05cb11d0293465
Signed-off-by: Miguel Perez <francisco.miguel.perez@ericsson.com>
Michael Vorburger [Mon, 25 Jul 2016 23:22:03 +0000 (01:22 +0200)]
Refactoring: MoreObjects2 is now in yangtools.utils
Change-Id: I2748b021c9c77d856b692445134beb70435f38d4
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Miguel Duarte [Wed, 5 Oct 2016 10:08:22 +0000 (12:08 +0200)]
Bug 6776 - Bad instructions returned by genius RPC
When calling the 'getEgressInstructionsForInterface' genius RPC, the
correct actions are returned, but encapsulated within a write-actions
deferred action set, which lead to a openflow plugin error.
Since these actions have to be acted upon immediately, and not at the
end of the pipeline, which is when an write-actions instruction will act
upon the actions, the code is modified to encapsulate the actions in an
apply-actions instruction, whose effects are immediate.
Change-Id: Icaedf148b015d729f21300209f84eeeb5b9b76dd
Signed-off-by: Miguel Duarte <miguel.duarte.de.mora.barroso@ericsson.com>
Sam Hague [Wed, 5 Oct 2016 12:22:24 +0000 (12:22 +0000)]
Merge "Adding the Add/Remove ExternalEndpoint commands."
David Suarez [Wed, 5 Oct 2016 10:05:49 +0000 (12:05 +0200)]
Make local variables creation and assignment in a single statement.
Some other minor formatting (removing commented code, etc.)
Change-Id: I009aaf17a4261a8f6f3147cbe3179f27253f773f
Signed-off-by: David Suarez <david.suarez.fuentes@ericsson.com>
Gobinath [Wed, 5 Oct 2016 07:12:16 +0000 (12:42 +0530)]
Arp cache feature changes
The aliveness monitor is used for the arp cache feature.
ARP monitor packets(ARP requests) are sent by the aliveness monitor and the
response to these are processed by the aliveness monitor.
The packets are checked whether they are from the L3 interface table. As
the ARP packets are now punted to controller from the L3 gateway mac table
(my mac changes), the packets are checked whether they come from that
table too.
Change-Id: Ic5ad946aac7a9f1a1210e55ba9ca39cb3c0865a6
Signed-off-by: Gobinath <gobinath@ericsson.com>
Faseela K [Wed, 5 Oct 2016 05:54:45 +0000 (05:54 +0000)]
Merge "Bug 6786: L3VPN is not honoring VTEP add or delete in operational cloud"
HANAMANTAGOUD V Kandagal [Thu, 22 Sep 2016 14:12:07 +0000 (19:42 +0530)]
Bug 6786: L3VPN is not honoring VTEP add or delete in operational cloud
Whenever VTEP IP on a DPN is added or deleted , the next-hop IP of VMs in
FIB entries on that DPN needs to be updated appropriately.
To know which all VM FIB entries to be modified , a new DpnToInterfaceList
operational datastore (DS) is being introduced in InterfaceManager.
InterfaceManager will update the DpnToInterfaceList operational DS , when
VM is booted or deleted.
Whenever VTEP IP is added/deleted on a DPN , VPN would get a VxLAN tunnel
event. VPN will query the InterfaceManager using RPC for a given DPNId to
retrieve the list of VpnInterfaces(VM ports).
Change-Id: Ia9fcc97712d6f2592d53ea98b106bbca0d2a1bf7
Signed-off-by: HANAMANTAGOUD V Kandagal <hanamantagoud.v.kandagal@ericsson.com>
Chintan Apte [Tue, 23 Aug 2016 05:57:06 +0000 (11:27 +0530)]
Adding the Add/Remove ExternalEndpoint commands.
Change-Id: I32f3c611a2641126c867ab4bc36423e647a41765
Signed-off-by: Chintan Apte <chintan.apte@ericsson.com>
Faseela K [Tue, 4 Oct 2016 09:51:40 +0000 (09:51 +0000)]
Merge " changes for REST API to make TZ specific parameters (transport zone name, tunnel type) mandatory."
Suraj Ranjan [Tue, 4 Oct 2016 05:57:17 +0000 (11:27 +0530)]
BUG 6838: Retry Mechanism for Batched Transaction
If the existing batch mechanism fails, then the individual operations of
each of the transactions will be retried so that all other operations
other than the problematic one will succeed.
Change-Id: Iec94654f1e28d9c9088cb0a6e3c6f7dece41e138
Signed-off-by: Suraj Ranjan <suraj.ranjan@ericsson.com>
Faseela K [Tue, 4 Oct 2016 05:33:03 +0000 (05:33 +0000)]
Merge "Bug 6642 - Improvising Batching code"
Sam Hague [Mon, 3 Oct 2016 23:57:35 +0000 (23:57 +0000)]
Merge "Add fcapsapplication-impl XML config to features/pom.xml"
Sam Hague [Mon, 3 Oct 2016 23:56:45 +0000 (23:56 +0000)]
Merge "Implement an action type nx_load_in_port"
Sam Hague [Mon, 3 Oct 2016 23:53:38 +0000 (23:53 +0000)]
Merge "Fixing overflow in long-to-IPv4 address conversion"
Sam Hague [Mon, 3 Oct 2016 20:29:35 +0000 (20:29 +0000)]
Merge "Revert "BUG 6838: Retry Mechanism for Batched Transaction""
Sam Hague [Mon, 3 Oct 2016 19:39:39 +0000 (19:39 +0000)]
Revert "BUG 6838: Retry Mechanism for Batched Transaction"
The corresponding netvirt patch is fairly large and conflicting with many other patches, so we will revert this patch until we work through the backlog.
This reverts commit
29f20d962ea8c3e76b4e81857f39b7d56cbe8769.
Change-Id: I536f9c75eeadf2bd9f3a622591ef386aa37914d8
Author: Sam Hague <shague@redhat.com>
Signed-off-by: Sam Hague <shague@redhat.com>
K.V Suneelu Verma [Wed, 31 Aug 2016 05:07:09 +0000 (10:37 +0530)]
Bug 6589 adding support for hwvtep devices ha
Added the following utility methods
to add a node as ha child node of the other node
to check if a node is ha child node or not
to get list of child nodes for given parent node
When two southbound devices are in mlag, the northbound l2gw configuration
has to be synced in both the devices.
a node is marked as ha child node based on other_config columnn key ha_id in
manager table
Two ha devices will have same ha_id value.
Change-Id: I4e4d58074b8a4b7478fe7859dd436d6ca8cf2d37
Signed-off-by: K.V Suneelu Verma <k.v.suneelu.verma@ericsson.com>
Faseela K [Mon, 3 Oct 2016 08:37:24 +0000 (08:37 +0000)]
Merge "BUG 6838: Retry Mechanism for Batched Transaction"
Faseela K [Mon, 3 Oct 2016 07:33:58 +0000 (07:33 +0000)]
Merge "Replace some collection.size() > 0 for !collection.isEmpty() to improve readability. Some other minor changes."
Faseela K [Mon, 3 Oct 2016 07:25:32 +0000 (07:25 +0000)]
Merge "Add support to the ITM to create Transport Zones with different UDP: VxLAN: default port VxLAN-GPE: 4880"
Faseela K [Mon, 3 Oct 2016 07:19:23 +0000 (07:19 +0000)]
Merge "Add SFC relevant service binding constants"
Kiran N Upadhyaya [Thu, 25 Aug 2016 13:52:44 +0000 (19:22 +0530)]
BUG 6838: Retry Mechanism for Batched Transaction
If the existing batch mechanism fails, then the individual operations of
each of the transactions will be retried so that all other operations
other than the problematic one will succeed.
Change-Id: I7ad598fe9604aae9ad1df71270f1cba5b212a634
Signed-off-by: Kiran N Upadhyaya <kiran.n.upadhyaya@ericsson.com>
David Suarez [Thu, 29 Sep 2016 16:50:17 +0000 (18:50 +0200)]
Replace some collection.size() > 0 for !collection.isEmpty() to
improve readability. Some other minor changes.
Change-Id: I2f41632125121fb3d2722dd5345f7e5f12525574
Signed-off-by: David Suarez <david.suarez.fuentes@ericsson.com>
Jaime Caamaño Ruiz [Thu, 22 Sep 2016 15:49:19 +0000 (17:49 +0200)]
Add SFC relevant service binding constants
Using the same service index as SCF, as they should not coexist, to
avoid shifting the following indexes.
Change-Id: I005c855372fdf62df32e615fb2fd16deed562c91
Signed-off-by: Jaime Caamaño Ruiz <jaime.caamano.ruiz@ericsson.com>
Marcos López Samaniego [Wed, 28 Sep 2016 15:54:25 +0000 (17:54 +0200)]
Fix Improved error message for jobs
Change-Id: I96f04654089ef390bfc2bab3c07e48f854254e3d
Signed-off-by: Marcos López Samaniego <marcos.lopez.samaniego@ericsson.com>
eraulhe [Wed, 28 Sep 2016 12:15:55 +0000 (14:15 +0200)]
Improved error message for jobs
Signed-off-by: eraulhe <raul.hernandez.sucuesta@ericsson.com>
Change-Id: Ia6942d70dd5186c9d19564ece507c08ca683bac0
Vratko Polak [Tue, 24 May 2016 15:15:16 +0000 (17:15 +0200)]
Add fcapsapplication-impl XML config to features/pom.xml
Otherwise standalone build of genius-features fails on SingleFeatureTest.
Change-Id: I76176cd7b7cab9d998e827612ee705df142faf11
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
David Suarez [Tue, 27 Sep 2016 15:19:32 +0000 (17:19 +0200)]
Add support to the ITM to create Transport Zones with different UDP:
VxLAN: default port
VxLAN-GPE: 4880
Signed-off-by: David Suarez <david.suarez.fuentes@ericsson.com>
Change-Id: I4bd746b9a144e957cf05664fbd1e3a32a1bcda09
Signed-off-by: David Suarez <david.suarez.fuentes@ericsson.com>
Olga Schukin [Tue, 27 Sep 2016 11:26:05 +0000 (14:26 +0300)]
Implement an action type nx_load_in_port
Implement action type: nx_load_in_port required for ping responder
Change-Id: I1629d809c28a2bf834f0aba5e960d47444d8c891
Signed-off-by: Olga Schukin <olga.schukin@hpe.com>
Yugandhar Reddy Kaku [Tue, 6 Sep 2016 10:53:58 +0000 (16:23 +0530)]
Bug 6642 - Improvising Batching code
Improvising Batching code :
A new concrete class(DefaultBatchHandler) should be written which will
Implement "ResourceHandler" and provide the functionality of Resource
Handler.
The current implementation requires modules to implement interface
"ResourceHandler" for using Batching functionality, with the new
implementation, this is not required. modules can register for batching
passing the object of DefaultBatchHandler.
The below methods are exposed in class "ResourceBatchingManager" for the
modules to use.
provide interfaces similar to databroker transaction interface.
public void put(String resourceType, InstanceIdentifier identifier,
DataObject updatedData);
public void merge(String resourceType, InstanceIdentifier identifier,
DataObject updatedData);
public void delete(String resourceType, InstanceIdentifier identifier);
Change-Id: I970df6a7351b4d432318c7ed1978be382e7eb0da
Signed-off-by: Yugandhar Reddy Kaku <yugandhar.reddy.kaku@ericsson.com>
Sam Hague [Thu, 22 Sep 2016 11:52:27 +0000 (11:52 +0000)]
Merge "Reverting Overriding in_port in table0 with Zero value"
xcheara [Thu, 22 Sep 2016 09:51:52 +0000 (15:21 +0530)]
Reverting Overriding in_port in table0 with Zero value
Description : It's been observed that the fix for Bug 6765 is resulting
flooding of ELAN packets when External IPs case.
Change-Id: Idfb5825439a993c7bc779454efc50436df188cd3
Signed-off-by: xcheara <chetan.arakere@altencalsoftlabs.com>
Georg Kunz [Wed, 21 Sep 2016 22:36:07 +0000 (00:36 +0200)]
Fixing overflow in long-to-IPv4 address conversion
This patch fixes two issues in the long-to-IPv4 address conversion
implemented in the MDSALUtils class:
* An integer signed/unsigned conversion overflow occuring when converting
IP with octets > 127. Results included IPs such as -64.-88.1.10 instead of
192.168.1.10.
* If the IP prefix mask provided to the conversion function was 0, it was
wrongly added to the IP address, such as 192.168.0.10/0. The resulting
OpenFlow mod flow message hence set the mask bit to 1 despite no actual
mask being set, resulting the switch rejecting the flow.
This patch is based on a fix for the old vpnservice project.
Change-Id: Ie64e8bdbdac7af656f36e814cb1d05a2c3431cd7
Signed-off-by: Georg Kunz <georg.kunz@ericsson.com>
Sam Hague [Wed, 21 Sep 2016 19:10:02 +0000 (19:10 +0000)]
Merge "Implement action types required for ping responder"
Michael Vorburger [Wed, 21 Sep 2016 13:04:29 +0000 (15:04 +0200)]
Add JavaDoc to AsyncDataTreeChangeListenerBase init() re. @PostConstruct
Clarification re. precise inheritance semantics of the JSR 250
@PostConstruct DI lifecycle annotation. Further details are also
documented on
https://wiki.opendaylight.org/view/BestPractices/DI_Guidelines#Nota_Bene
This investigate and clarification has come out of the problem
encountered and solved in
https://git.opendaylight.org/gerrit/#/c/45966/.
Change-Id: If7f62aa6818706943c3555c55b86b91b7233e66a
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Faseela K [Wed, 21 Sep 2016 06:12:53 +0000 (06:12 +0000)]
Merge "Bug 6765 : Overriding in_port in table0 with Zero value"
Faseela K [Wed, 21 Sep 2016 05:42:19 +0000 (05:42 +0000)]
Merge "Performs a residual cleanup of ElanPseudoPort flows"
xcheara [Fri, 9 Sep 2016 15:18:11 +0000 (20:48 +0530)]
Bug 6765 : Overriding in_port in table0 with Zero value
Description : OVS split-horizon logic will drop pkts by default if value
of in_port and out_port is same. But, if we have two vlan-trunk member
VM for which trunk interface is same, the connectivity between these two
VM cannot be established as OVS drops such packets. So, it is required to
override the value of in_port so that split_horizon check is bypassed.
The in_port value will be overrriden using load(ex: load:0->NXM_OF_IN_PORT[]).
Change-Id: Iab5b87558beecee76a011bad33e1867311dc4a29
Signed-off-by: xcheara <chetan.arakere@altencalsoftlabs.com>
Depends-On: Ica0a1b2da27fc189b8906462a4420473215a22af
Signed-off-by: xcheara <chetan.arakere@altencalsoftlabs.com>
Tomer Pearl [Mon, 19 Sep 2016 14:54:15 +0000 (17:54 +0300)]
Bug 6748: Added support for match on nxm_reg5
Change-Id: Ifb87cea2142e90d29fcb0859ab1c9afb265a8e47
Signed-off-by: Tomer Pearl <tomer.pearl@hpe.com>
Miguel Perez [Mon, 11 Jul 2016 11:21:36 +0000 (13:21 +0200)]
Performs a residual cleanup of ElanPseudoPort flows
+ When an ElanPseudoPort is removed on the Northbound, we
must ensure that no elanPseudoPort related flow is left
behind in any DPN.
Change-Id: I7201a428272b638d2055b82f7110a1b86fbad2fa
Signed-off-by: Miguel Perez <francisco.miguel.perez@ericsson.com>
Faseela K [Fri, 16 Sep 2016 18:06:27 +0000 (18:06 +0000)]
Merge "NPE in InterfaceTopologyStateListener"
Faseela K [Fri, 16 Sep 2016 18:05:29 +0000 (18:05 +0000)]
Merge "VM Migration: Flows not programmed in new DPN"