#opendaylight-meeting: kernel projects
Meeting started by rgoulding at 15:54:26 UTC
(full logs).
Meeting summary
- agenda bashing (rgoulding, 15:54:30)
- JavaDoc sites, next step? Merge
https://git.opendaylight.org/gerrit/#/c/60213/ (rgoulding,
15:54:38)
- continued from last week… Discussion: Should
each thing create its own ExecutorService or should we have a single
central shared one? Like an EE servers's manged thread pools.
(rgoulding,
15:54:51)
- Websocket and clustering:
https://git.opendaylight.org/gerrit/#/c/63264/ (rgoulding,
15:54:58)
- merged/pending patches in the last week
(rgoulding,
15:55:10)
- https://www.opendaylight.org/events/odl-oxygen-developer-design-forum
(rgoulding,
15:59:56)
- Oxygen DDF (rgoulding, 16:03:43)
- https://www.opendaylight.org/events/odl-oxygen-developer-design-forum
(rgoulding,
16:03:54)
- https://wiki.opendaylight.org/view/Events:Oxygen_Dev_Forum
(rgoulding,
16:04:56)
- JavaDoc site docs (rgoulding, 16:05:55)
- JavaDoc sites, next step? Merge
https://git.opendaylight.org/gerrit/#/c/60213/ (rgoulding,
16:06:01)
- https://git.opendaylight.org/gerrit/#/c/60213/
(rgoulding,
16:06:06)
- Michael V. feels stuck on how to move forward
with this. We probably need Thanh’s help (rgoulding,
16:06:29)
- instead of just putting it in “sites” put it in
a new location perhaps (rgoulding,
16:10:14)
- perhaps something like “/javadocs” (rgoulding,
16:10:21)
- that way we can have both systems concurrently
while making the transition (rgoulding,
16:10:33)
- vorburger feels like this is caution for
cuation’s sake because he doesn’t know if there are any users of
this (rgoulding,
16:10:48)
- each component had its own maven site, which
made it somewhat hard to use (rgoulding,
16:12:05)
- but the point is there could have been some
users (rgoulding,
16:12:11)
- if we determine that we can build javadoc in
the new location w/ better format then we can probably stop building
theold one (rgoulding,
16:12:53)
- ACTION: vorburger to
open helpdesk ticket to make anew site for “javadoc(s)" (rgoulding,
16:13:16)
- we need the javadocs to be deployed to some
directory, then we could point the site to the directory
(rgoulding,
16:14:10)
- vorburger says it is put in target (rgoulding,
16:14:21)
- deploy-site.xml example https://github.com/opendaylight/odlparent/blob/master/deploy-site.xml
(zxiiro,
16:15:43)
- https://github.com/opendaylight/odlparent/blob/master/deploy-site.xml#L38
(rgoulding,
16:16:06)
- we could put the javadocs here, or change this
to wherever the javadoc is generated (rgoulding,
16:16:40)
- https://github.com/opendaylight/odlparent/blob/master/deploy-site.xml#L20
(rgoulding,
16:17:21)
- content/sites/javadoc instead of
content/sites/site for the second site url (rgoulding,
16:17:39)
- every project will need this file to
deploy (rgoulding,
16:18:04)
- we could get jjb to push the file in
(rgoulding,
16:18:14)
- start with odlparent first then figure out how
to handle other projects (rgoulding,
16:18:26)
- Websocket and clustering: https://git.opendaylight.org/gerrit/#/c/63264/ (rgoulding, 16:19:07)
- https://git.opendaylight.org/gerrit/#/c/63264/
(rgoulding,
16:19:19)
- https://bugs.opendaylight.org/show_bug.cgi?id=9147
(rgoulding,
16:19:46)
- https://git.opendaylight.org/gerrit/#/q/topic:ListenerAdapterTreeMigration
(rgoulding,
16:20:39)
- it appears that ths is the right
approach (rgoulding,
16:23:33)
- this code just predates that change
(rgoulding,
16:23:37)
- ClusteredDOMDataTreeChangeListener should
eventually be killed eventually, but probably not as relevant to
this individual use case (rgoulding,
16:24:51)
- it is mroe forward looking (rgoulding,
16:24:56)
- tree based change is by design (rgoulding,
16:26:52)
- Discussion: Should each thing create its own ExecutorService or should we have a single central shared one? Like an EE servers's manged thread pools. (rgoulding, 16:29:58)
- https://git.opendaylight.org/gerrit/#/c/59177/
(rgoulding,
16:30:17)
- prompted by thinking about how
https://git.opendaylight.org/gerrit/#/c/59177/ users should obtain
an ExecutorService (rgoulding,
16:30:27)
- https://bugs.opendaylight.org/show_bug.cgi?id=11
(rgoulding,
16:30:40)
- one application may starve another if there is
not application awareness in the threading model (rgoulding,
16:32:41)
- https://git.opendaylight.org/gerrit/#/c/63439/
(vorburger,
16:32:44)
- the existing mechanisms are not meant to be
global services,they are meant for morelocal use cases (rgoulding,
16:33:22)
- threadpools are allocated to handle a certain
exchange point (i.e., DTCL) (rgoulding,
16:33:38)
- so they are grouped by workflow (rgoulding,
16:33:44)
- there is a competing bug that says we should be
able to better isolate application workflows (rgoulding,
16:34:04)
- https://git.opendaylight.org/gerrit/#/c/63439
(rgoulding,
16:34:38)
- in this patch, this ensures that no work
stealing is achieved. each workload is mapped to a key (rgoulding,
16:35:28)
- so you don’t have a Thread per DTCL
(rgoulding,
16:35:56)
- merged/pending patches (rgoulding, 16:36:52)
- https://git.opendaylight.org/gerrit/#/c/48710/
(rgoulding,
16:37:51)
- not completely not impactful (rgoulding,
16:37:57)
- David is putting in additional changes to
adjust for this ^^ (rgoulding,
16:38:08)
- does this impact only trailing whitespace in
src/main/resources (rgoulding,
16:38:19)
- activates checkstyle on src/main/resources, but
only real impact is that non-java files are validated (line endings,
no trailing tabs/spaces) (rgoulding,
16:38:53)
- is checkstyle checking the extension?
(rgoulding,
16:39:44)
- vorburger thinks just by file extension
(rgoulding,
16:39:51)
- when odlparent with this patch is adopted there
may be some work to clean up src/main/resources (rgoulding,
16:40:23)
- should be trivial and there is info on how to
make the change when we transition to the new odlparent
version (rgoulding,
16:40:38)
- no one really barked about putting this
in (rgoulding,
16:41:05)
- this will need to be release noted for sure
though (rgoulding,
16:41:11)
- in 3.0.0 (rgoulding,
16:41:15)
- this patch is listed in an email that vorburger
sent out and agreed we should handle this with release note
(rgoulding,
16:41:35)
- rovarga thinks this is okay; need to follow up
with Stephen who has -1 (rgoulding,
16:42:37)
- I'll remove my -1, and merge it (rgoulding,
16:42:55)
Meeting ended at 16:44:22 UTC
(full logs).
Action items
- vorburger to open helpdesk ticket to make anew site for “javadoc(s)"
Action items, by person
- vorburger
- vorburger to open helpdesk ticket to make anew site for “javadoc(s)"
People present (lines said)
- rgoulding (71)
- zxiiro (6)
- odl_meetbot (3)
- vorburger (2)
Generated by MeetBot 0.1.4.