[NETCONF-96] Standard ietf-netconf-monitoring.yang file has been modified to import specific versions of dependencies that could break other mounted devices Created: 04/Nov/15  Updated: 15/Mar/19  Resolved: 19/Apr/16

Status: Resolved
Project: netconf
Component/s: netconf
Affects Version/s: None
Fix Version/s: None

Type: Bug
Reporter: Balaji Varadaraju Assignee: Ryan Goulding
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All

Issue Links:
Duplicate
is duplicated by NETCONF-92 changes to notifications yang models ... Resolved
is duplicated by NETCONF-120 Allow cache/schema to be configured f... Resolved
External issue ID: 4577
Priority: High

 Description   

Standard ietf-netconf-monitoring.yang has been modified to import specific versions of ietf-yang-types and ietf-inet-types. This is a deviation from standard IETF file. I've gone through the history of this file and it is suggested that this was done to load classes properly.

However this will break other mounted devices that uses the standard version of this file and expect that latest versions of the import modules will be used.

I removed the explicit version from imports, restarted the controller and it worked fine for both controller netconf, and other mounted netconf devices.

So I think we need to look into reverting these changes in the IETF-NETCONF-MONITORING.yang file and maintain the standard IETF version so that it can play well with other devices.

 Comments   
Comment by Balaji Varadaraju [ 11/Nov/15 ]

Gerrit Patch URL

https://git.opendaylight.org/gerrit/#/c/29566/1

Comment by Robert Varga [ 13/Nov/15 ]

Move to NETCONFI project.

Comment by Balaji Varadaraju [ 13/Nov/15 ]

WE need this fixed in stable lithium. That's where I've pushed the fix.

Comment by Ryan Goulding [ 16/Nov/15 ]

Hi Tony,

Was

Comment by Ryan Goulding [ 16/Nov/15 ]

Tony & Maros, can you take a look at the patch Bala has created? Please let me know your thoughts so we can come up with a solution for this. I'll bring it up in the weekly MD-SAL call tomorrow as well.

Thanks,
Ryan

Comment by Ryan Goulding [ 24/Nov/15 ]

Do we have a survey of who expects 2010 vs 2013 model? This seems to be affecting multiple groups of people.

Comment by Robert Varga [ 24/Nov/15 ]

Another way of dealing with this would be to make the cache directory configurable. The cache directory is currently hard-coded in opendaylight/netconf/sal-netconf-connector/src/main/java/org/opendaylight/controller/config/yang/md/sal/connector/netconf/NetconfConnectorModule.java:218.

It would be nice to be able to specify this cache directory on a per-device basis, so that conflicting YANG texts do not end up poisoning each other.

Comment by Giles Heron [ 24/Nov/15 ]

have you been speaking to Tony, Robert? That's exactly what I was just suggesting to him

Comment by Ryan Goulding [ 01/Dec/15 ]

Draft patch still working on testing:
https://git.opendaylight.org/gerrit/#/c/30163/

Comment by Ryan Goulding [ 06/Jan/16 ]

Lithium patch is ready for review.

Comment by Ryan Goulding [ 19/Jan/16 ]

stable/beryllium -> https://git.opendaylight.org/gerrit/#/c/33013/
master -> https://git.opendaylight.org/gerrit/#/c/32362/

Comment by Ryan Goulding [ 19/Jan/16 ]

Wiki Documentation Added:
https://wiki.opendaylight.org/view/OpenDaylight_Controller:Config:Examples:Netconf#Aggregation_of_a_separate_SharedSchemaCache.2C_SchemaContextFactory_and_FilesystemSchemaSourceCache

Comment by Alexis de Talhouët [ 05/Feb/16 ]

First iteration done here https://git.opendaylight.org/gerrit/#/c/30163 introduce a BUG: sideloaded schemas, manually added in cache/schema, and specified in the netconf connector, where not being taking in account in the specified cache folder. This patch make sure sideloaded schemas are copied in the specified directory, and make sure those schemas are correctly processed by FilesystemSchemaSourceCache. The reason why we create the directory and add the schemas in there before create the instance of FilesystemSchemaSourceCache is because FilesystemSchemaSourceCache loads the schemas from the dir only once, during the instantiation.

lithium --> https://git.opendaylight.org/gerrit/#/c/34156/

Comment by Alexis de Talhouët [ 05/Feb/16 ]

stable/beryllium --> https://git.opendaylight.org/gerrit/#/c/34189/
master --> https://git.opendaylight.org/gerrit/#/c/34188/

Generated at Wed Feb 07 20:14:10 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.