[BGPCEP-311] New openconfig modules frequently cause "server is unhealthy". Created: 13/Nov/15 Updated: 03/Mar/19 Resolved: 20/Nov/15 |
|
| Status: | Resolved |
| Project: | bgpcep |
| Component/s: | BGP |
| Affects Version/s: | Bugzilla Migration |
| Fix Version/s: | Bugzilla Migration |
| Type: | Bug | ||
| Reporter: | Vratko Polak | Assignee: | Claudio David Gasparini |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Environment: |
Operating System: All |
||
| Issue Links: |
|
||||||||
| External issue ID: | 4633 | ||||||||
| Description |
|
Change 29551 [0] which fixed 2015-11-13 04:35:19,066 | WARN | config-pusher | JavassistUtils | 62 - org.opendaylight.mdsal.binding-generator-impl - 0.8.0.SNAPSHOT | Failed to customize org.opendaylight.yang.gen.v1.http.openconfig.net.yang.bgp.rev151009.bgp.global.base.Config$StreamWriter from prototype org.opendaylight.yangtools.binding.data.codec.gen.impl.DataObjectSerializerPrototype javassist.CannotCompileException: [source error] no such class: org.opendaylight.yang.gen.v1.urn.ietf.params.xml.ns.yang.ietf.inet.types.rev130715.AsNumber at javassist.CtBehavior.setBody(CtBehavior.java:446) Specific reason is not clear to me. Maybe the new BGP models end up using incorrect revision of ietf-inet-types, or perhaps MDSAL does not expose the newer revision correctly. |
| Comments |
| Comment by Vratko Polak [ 13/Nov/15 ] |
|
Downstream consequences: Merges to integration/testing get -1 from verify job. Merges to integration/distribution get -1 from verify job. |
| Comment by Claudio David Gasparini [ 13/Nov/15 ] |
| Comment by Vratko Polak [ 14/Nov/15 ] |
|
> Another alternative is to bump yang-types/ietf-types across the board – not sure how feasible that is. I tried, not sure if it was worth of the time I spent on it. |
| Comment by Vratko Polak [ 14/Nov/15 ] |
|
I do not see an easy way to generate .yang.network.topology.rev131021. classes that would work with both older .yang.ietf.inet.types.rev100924. classes and the newer .yang.ietf.inet.types.rev130715. ones. Aside from just pasting contents of rev130715 yang file into the rev100924 one and keep using that (changes are compatible this way). I do not expect the long list of project to get their version of Change 29692. The yang model pasting solution may be the easiest one, followed by removing openconfig stuff from Beryllium Bgpcep. Do we plan some serious design effort on how to handle multiple revisions and revision bumping in Boron? |
| Comment by Vratko Polak [ 14/Nov/15 ] |
|
> The yang model pasting solution may be the easiest one, Just to be clear, openconfig removal is a separate solution, not a next step of the first solution. By the way, we now have new job for system testing multiple patches at once. Change 29692 and Change 29660 fared like this: Bgpcep test job failed: https://jenkins.opendaylight.org/sandbox/job/bgpcep-csit-1node-userfeatures-all-beryllium/3/console 2015-11-14 10:20:01,941 | ERROR | config-pusher | ConfigRegistryImpl | 88 - org.opendaylight.controller.config-manager - 0.4.0.SNAPSHOT | Configuration Transaction failed on 2PC, server is unhealthy |
| Comment by Vratko Polak [ 18/Nov/15 ] |
|
My testing suggests this patch by Claudio has no downsides. |