| sousaedu | 6774898 | 2021-03-23 01:02:02 +0100 | [diff] [blame] | 1 | ####################################################################################### |
| 2 | # Copyright ETSI Contributors and Others. |
| 3 | # |
| 4 | # Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | # you may not use this file except in compliance with the License. |
| 6 | # You may obtain a copy of the License at |
| 7 | # |
| 8 | # http://www.apache.org/licenses/LICENSE-2.0 |
| 9 | # |
| 10 | # Unless required by applicable law or agreed to in writing, software |
| 11 | # distributed under the License is distributed on an "AS IS" BASIS, |
| 12 | # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or |
| 13 | # implied. |
| 14 | # See the License for the specific language governing permissions and |
| 15 | # limitations under the License. |
| 16 | ####################################################################################### |
| 17 | --- |
| 18 | deprecations: |
| 19 | - | |
| 20 | OSM Information Model (IM) Debian package is no longer required, therefore |
| 21 | its usage has been removed. |
| 22 | security: |
| 23 | - | |
| 24 | Security scanning has been added through the usage of the Gitlab scanners. |
| 25 | fixes: |
| 26 | - | |
| 27 | Fixes [bug 1407](https://osm.etsi.org/bugzilla/show_bug.cgi?id=1407). This |
| 28 | bug was introduced due to the change from the old descriptor models into |
| 29 | the new SOL006 descriptors. The Resource Orchestrator (RO) didn't received |
| 30 | the information relating to port security, therefore it couldn't determine |
| 31 | if the network interface should have it enabled or disabled, which in turn |
| 32 | always applied the default for the used substrate. |
| 33 | - | |
| 34 | Fixes [bug 1408](https://osm.etsi.org/bugzilla/show_bug.cgi?id=1408). This |
| 35 | bug was introduced due to the change from the old descriptor models into |
| 36 | the new SOL006 descriptors. The Resource Orchestrator (RO) didn't received |
| 37 | the information relating to Virtual Link Descriptor (VLD) IP Profile, |
| 38 | therefore it couldn't apply the desired values for subnets and DHCP. |
| 39 | - | |
| 40 | Fixes [bug 1446](https://osm.etsi.org/bugzilla/show_bug.cgi?id=1446). This |
| 41 | bug was introduced due to the change from the old Resource Orchestrator to |
| 42 | the new Next-Generation Resource Orchestrator. For some Network Services, |
| 43 | we have observed that sometimes the NS remains in state BUILDING. Checking |
| 44 | in the database for ro_task we have seen that this happens because the RO |
| 45 | keeps on locking always the same ro_task that has dependencies on another |
| 46 | ro_task that is never executed. |
| 47 | - | |
| 48 | Fixes [bug 1445](https://osm.etsi.org/bugzilla/show_bug.cgi?id=1445). This |
| 49 | bug was introduced due to the change from the old Resource Orchestrator to |
| 50 | the new Next-Generation Resource Orchestrator. When we try to instantiate a |
| 51 | service with SRIOV interfaces it fails. Debugging the error we have seen |
| 52 | that the mac does not arrive to the onosof sdn plugin. |
| 53 | - | |
| 54 | Fixes [bug 1386](https://osm.etsi.org/bugzilla/show_bug.cgi?id=1386). This |
| 55 | bug was introduced due to the change from the old Resource Orchestrator to |
| 56 | the new Next-Generation Resource Orchestrator. RO has a database locking |
| 57 | system intended for HA. Every worker locks a "ro_task" by writting at |
| 58 | database, before performing it at the VIM, and unlocks afterwards. In case |
| 59 | of a crash (e.g. POD is destroyed) the ro_task becames locked during many |
| 60 | time until lock expires. The expiration time is set to 1 hour as this is |
| 61 | the time that can take some ro_task (some VMs can take this time). It will |
| 62 | be better having a shorter expiration time, and re-new it when needed if |
| 63 | the thread that perform the VIM task is still alive. In that a way after a |
| 64 | crash, new RO POD will take over the locked ro_task quite soon. The change |
| 65 | has a new asyncio task that performs the renew of the locks of all the |
| 66 | needed vim_threads. Vim_thread cannot do it by themselves because some VIM |
| 67 | plugin operations are synchronous. |
| 68 | - | |
| 69 | Fixes [bug 1385](https://osm.etsi.org/bugzilla/show_bug.cgi?id=1385). This |
| 70 | bug was introduced due to the change from the old Resource Orchestrator to |
| 71 | the new Next-Generation Resource Orchestrator. Code creates vim_threads |
| 72 | when there is new VIMs to handle (NS deployed at this VIM). However when |
| 73 | the VIM is not used anymore (NS is deleted), vim_thread is not re-used in |
| 74 | the next assignation. This is something that can be easily see with the |
| 75 | logs. This bug does not causes malfunction, but can lead to have some |
| 76 | threads overloaded. In addition resources are not free when VIM is not |
| 77 | used. |
| 78 | other: |
| 79 | - | |
| 80 | The code has been reformatted to be compliant with the Python Language |
| 81 | standards, using [black](https://github.com/psf/black) and |
| 82 | [flake8](https://flake8.pycqa.org/en/latest/). |