2019-07-24 05:46:49 UTC - Adam Varsano: Hi,
If I'll upload the go code without compile them?
then we can add prewarm actions, should I do that?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563947209100900?thread_ts=1563895719.099200&cid=C3TPCAQG1
----
2019-07-24 05:49:33 UTC - chetanm: Can you drop a mail on
<mailto:dev@openwhisk.apache.org|dev@openwhisk.apache.org> with details. Then
others can also share there thoughts on psosible solutions
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563947373101100?thread_ts=1563895719.099200&cid=C3TPCAQG1
----
2019-07-24 10:06:45 UTC - Adam Varsano: sure
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563962805101500?thread_ts=1563895719.099200&cid=C3TPCAQG1
----
2019-07-24 11:15:10 UTC - Roberto Santiago: Seeing some intermittent problems
on my k8s install. Looked into the logs of the pods and found this on the
CouchDB logs. We are running it in k8s versus externally as is suggested.
None the less I am trying to understand what this error means:
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563966910103100
----
2019-07-24 11:15:12 UTC - Roberto Santiago: ```[error]
2019-07-24T11:07:10.454399Z couchdb@couchdb0 <0.263.0> -------- Could not
get design docs for
<<"shards/a0000000-bfffffff/verifytestdb.1561574453">>
error:{database_does_not_exist,[{mem3_shards,load_shards_from_db,"verifytestdb",[{file,"src/mem3_shards.erl"},{line,395}]},{mem3_shards,load_shards_from_disk,1,[{file,"src/mem3_shards.erl"},{line,370}]},{mem3_shards,for_db,2,[{file,"src/mem3_shards.erl"},{line,54}]},{fabric_view_all_docs,go,5,[{file,"src/fabric_view_all_docs.erl"},{line,24}]},{couch_db,'-get_design_docs/1-fun-0-',1,[{file,"src/couch_db.erl"},{line,627}]}]}
[error] 2019-07-24T11:07:10.454609Z couchdb@couchdb0 emulator -------- Error in
process <0.22511.384> on node couchdb@couchdb0 with exit value:
{database_does_not_exist,[{mem3_shards,load_shards_from_db,"verifytestdb",[{file,"src/mem3_shards.erl"},{line,395}]},{mem3_shards,load_shards_from_disk,1,[{file,"src/mem3_shards.erl"},{line,370}]},{mem3_shards,for_db,2,[{file,"src/mem3_shards.erl"},{line,54}]},{fabric_view_all_docs,go,5,[{file,"src/fabric_view_all_docs.erl"},{line,24}]},{couch_db,'-get_design_docs/1-fun-0-',1,[{file,"src/couch_db.erl"},{line,627}]}]}```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563966912103300?thread_ts=1563966912.103300&cid=C3TPCAQG1
----
2019-07-24 11:20:22 UTC - Satwik Kolhe: I had issues with CouchDB, where the PV
was provisioned using Rook. Did you check if there is disk space available for
the couchdb pod?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563967222103400?thread_ts=1563966912.103300&cid=C3TPCAQG1
----
2019-07-24 11:24:52 UTC - Roberto Santiago: How would I check that?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563967492103600?thread_ts=1563966912.103300&cid=C3TPCAQG1
----
2019-07-24 11:27:08 UTC - Satwik Kolhe: kubectl exec -it
<ow-couchdb-pod-name> -n <ow-namespace> -- df -hT
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563967628103800?thread_ts=1563966912.103300&cid=C3TPCAQG1
----
2019-07-24 11:27:45 UTC - Roberto Santiago: ```Filesystem Size Used Avail
Use% Mounted on
overlay 102G 70G 32G 70% /
tmpfs 68M 0 68M 0% /dev
tmpfs 4.0G 0 4.0G 0% /sys/fs/cgroup
/dev/sda1 102G 70G 32G 70% /etc/hosts
shm 68M 0 68M 0% /dev/shm
/dev/sdc 136G 48G 88G 35% /opt/couchdb/data
tmpfs 4.0G 13k 4.0G 1% /run/secrets/kubernetes.io/serviceaccount
tmpfs 4.0G 0 4.0G 0% /proc/acpi
tmpfs 4.0G 0 4.0G 0% /proc/scsi
tmpfs 4.0G 0 4.0G 0% /sys/firmware```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563967665104000?thread_ts=1563966912.103300&cid=C3TPCAQG1
----
2019-07-24 11:29:27 UTC - Satwik Kolhe: Ok - Disk space is not an issue!
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563967767104200?thread_ts=1563966912.103300&cid=C3TPCAQG1
----
2019-07-24 11:34:21 UTC - Roberto Santiago: Along the same lines as my previous
message. I am seeing this in logs from the controller:
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968061105000
----
2019-07-24 11:34:24 UTC - Roberto Santiago: ```[2019-07-24T11:23:40.288Z]
[INFO] [#tid_sid_invokerHealth] [InvokerPool] invoker status changed to 0 ->
Offline, 1 -> Offline, 2 -> Offline, 3 -> Offline, 4 -> Offline, 5
-> Offline, 6 -> Offline, 7 -> Offline, 8 -> Offline, 9 ->
Offline, 10 -> Offline, 11 -> Offline, 12 -> Offline, 13 ->
Offline, 14 -> Offline, 15 -> Healthy, 16 -> Unresponsive, 17 ->
Healthy```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968064105200?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 11:35:01 UTC - Roberto Santiago: Is this normal? Does offline
simply mean that it is in a cold state ?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968101105800
----
2019-07-24 11:35:12 UTC - Roberto Santiago: How many invokers to people
typically use?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968112106100
----
2019-07-24 11:36:23 UTC - chetanm: offline means that no ping was received from
the Invoker
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968183106200?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 11:36:38 UTC - chetanm: So either no invoker pod exist
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968198106400?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 11:37:52 UTC - Roberto Santiago: I only have 3 invoker pods which I
guess are 15, 16 and 17. I was just confused with the presence of 0 through 14.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968272106600?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 11:38:14 UTC - Roberto Santiago: Should it be like that or is that a
symptom of something not working correctly
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968294106800?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 11:39:14 UTC - chetanm: I think if any invoker pod gets killed then
k8s starts a new invoker pod and for some reason its giving it a newer index
instead of reusing the index for any older pod
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968354107000?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 11:40:42 UTC - Roberto Santiago: ah! ok.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968442107200?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 11:42:42 UTC - Roberto Santiago: So a deleted the controller pod and
when it came back its still using those last three slots
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968562107400?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 11:43:53 UTC - Roberto Santiago: btw, how many invokers do you use?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563968633107600?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 11:57:32 UTC - chetanm: We are not using k8s but using Mesos. So
there its a different approach
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563969452107800?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 12:02:00 UTC - Roberto Santiago: Do you like it?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563969720108000?thread_ts=1563968064.105200&cid=C3TPCAQG1
----
2019-07-24 12:46:13 UTC - Sven Lange-Last: when invokers run on kube, they are
supposed to ask zookeeper for an index. the expected behaviour is that this
approach keeps the the index stable because it is always using the same key
over and over again when asking zookeeper for an index.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972373110100
----
2019-07-24 12:46:21 UTC - Sven Lange-Last: ^^ @Roberto Santiago
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972381110400
----
2019-07-24 12:48:00 UTC - Sven Lange-Last: when an invoker sends a ping message
to the controller, it includes its index. so controller checks whether it
already knows this index. if not, it extends the invoker array and fills the
gaps with offline invokers with the expectation that the invokers with unknown
indices will send their ping messages soon.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972480112300
----
2019-07-24 12:48:43 UTC - Sven Lange-Last: the current load balancer design
expects that the invoker array has no gaps and uses the indices sent by
invokers.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972523113200
----
2019-07-24 12:49:06 UTC - Sven Lange-Last: btw - large providers use hundreds
of invokers.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972546113600
----
2019-07-24 12:49:58 UTC - Sven Lange-Last: offline means that the controller
did not receive a ping message from an invoker with that index yet or recently.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972598114600
----
2019-07-24 12:50:44 UTC - Sven Lange-Last: unresponsive means that out of the
10 most recent invocations processed by the invoker, 3 did not complete in time.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972644115300
----
2019-07-24 12:51:28 UTC - Sven Lange-Last: in time means 2 * max(action time
limit, 60 seconds) + 1 minute.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972688116000
----
2019-07-24 12:52:22 UTC - Sven Lange-Last:
<https://github.com/apache/incubator-openwhisk/blob/4a09c73bb391fcc8b3cf0ecd76ac7abad7fe849d/core/controller/src/main/scala/org/apache/openwhisk/core/loadBalancer/CommonLoadBalancer.scala#L103>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972742116800
----
2019-07-24 12:53:30 UTC - Roberto Santiago: Ok. This makes a lot of sense. I
wonder why zookeeper keeps starting with such a high index. About 30 minutes
ago a deleted all pods and when everything came back to life I still end up
having indexes 15, 16 and 17 for my three invokers.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972810117800
----
2019-07-24 12:54:24 UTC - Roberto Santiago: Thank so much for the information.
I've been slowly making my way through all the code to learn all the pieces
really well. So I appreciate all these details.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972864119300
----
2019-07-24 12:54:54 UTC - Sven Lange-Last: if you do not remove zookeeper pod
or zookeeper uses a persistent volume for its data, it will keep data. the
question is which data is used in your kube cluster as a key for zookeeper when
assigning the id.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972894119800
----
2019-07-24 12:55:08 UTC - Sven Lange-Last: search for messages from
<https://github.com/apache/incubator-openwhisk/blob/master/core/invoker/src/main/scala/org/apache/openwhisk/core/invoker/InstanceIdAssigner.scala>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563972908120200
----
2019-07-24 12:57:39 UTC - Sven Lange-Last: @Roberto Santiago check this part:
<https://github.com/apache/incubator-openwhisk/blob/4a09c73bb391fcc8b3cf0ecd76ac7abad7fe849d/core/invoker/src/main/scala/org/apache/openwhisk/core/invoker/Invoker.scala#L107-L142>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563973059120800
----
2019-07-24 12:59:23 UTC - Roberto Santiago: very cool!
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563973163121100
----
2019-07-24 12:59:36 UTC - Roberto Santiago: Any insights about the couchDB
error I posted above?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563973176121400
----
2019-07-24 13:01:49 UTC - Sven Lange-Last: @Roberto Santiago sorry, i’m no
couchdb expert. i’m working at ibm and usually use ibm’s cloudant service.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563973309122200
----
2019-07-24 13:03:16 UTC - Roberto Santiago: No worries. This has been of great
help! I hope to be useful enough to be contributing code soon.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563973396122700
----
2019-07-24 13:04:24 UTC - Sven Lange-Last: i’m looking forward to your
contributions :slightly_smiling_face: do you already have anything in mind? any
areas where you need changes?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563973464123700
----
2019-07-24 13:07:31 UTC - chetanm: Thanks @Sven Lange-Last for details. Somehow
I was under the impression that for kube case with Docker CF we are using
ordinal number of the DaemonSet as id. But looks like we use the Zookeeper to
determine the id and use pod name as unique name
+1 : Sven Lange-Last
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563973651125400
----
2019-07-24 13:20:21 UTC - Roberto Santiago: @Sven Lange-Last I have a colleague
who is a kafka genius and very interested in openwhisk. We both share a
"vision" of stream based computing. When he took a look at how kafka is being
used he speculated that there was much more OW could be getting from kafka
especially for integrating into existing services that are kafka based. So no
particular feature request there just getting enough knowledge to analyze
intelligently. My particular focus is around scaling AI and compositional
architectures. Right now my most immediate focus is on very practical things
with OW (e.g. measuring/increasing performance, measuring/increasing
reliability, satisfying High Availability requirements, etc.). So right now,
just focusing on being the student to OW before proposing anything. If
anything my hope is to be sufficient for fixing bugs in the next few weeks.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563974421133000
----
2019-07-24 13:47:40 UTC - chetanm: @Roberto Santiago You can get the zookeeper
state via `kubectl exec <release>-zookeeper-0 zkCli.sh ls
/invokers/idAssignment/mapping`
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563976060133700
----
2019-07-24 13:48:13 UTC - chetanm: That should show the children under the
mapping path which would have node names
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563976093134400
----
2019-07-24 14:07:06 UTC - Matt Rutkowski: REMINDER: ~ 1 HOUR until the Apache
OpenWhisk Tech. Interchange project meeting,
* Day-Time: Every other Wednesday, 11AM EDT (Eastern US), 5PM CEST (Central
Europe), 3PM GMT, 11PM (Beijing)
* Zoom: <https://zoom.us/my/asfopenwhisk>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563977226134800
----
2019-07-24 14:33:49 UTC - chetanm: How does the api gateway support in OpenWhisk
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563978829135300
----
2019-07-24 14:34:15 UTC - chetanm: I see under route management 3 actions for
create/delete/get
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563978855135800
----
2019-07-24 14:35:03 UTC - chetanm: and there is an apigateway docker image
which probably needs Redis (and possibly minio for object storage)
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563978903136600
----
2019-07-24 14:37:58 UTC - chetanm: So when I do `wsk api create /hello /world
get hello` what would the action do? From code it appears there are couple of
lua scripts which support the CRUD api and there is a datastore abstraction
backed by Redis.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563979078138200
----
2019-07-24 14:38:18 UTC - chetanm: So upon api create it would possibly record
that info in redis?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563979098138800
----
2019-07-24 14:38:41 UTC - chetanm: How does it use that info for actual HTTP
calls?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563979121139300?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:10:13 UTC - Matt Rutkowski:
<https://thenewstack.io/the-openwhisk-serverless-platform-now-an-apache-top-level-project-builds-on-kubernetes-mesos/>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563984613139700
----
2019-07-24 16:10:23 UTC - Matt Rutkowski:
<https://devops.com/asf-moves-openwhisk-serverless-computing-forward/>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563984623140000
----
2019-07-24 16:10:32 UTC - Matt Rutkowski:
<https://it.toolbox.com/blogs/shrutiumathe/apache-openwhisk-community-graduates-to-become-an-apache-software-foundation-asf-top-level-project-072419>
partyparrot : Michael Schmidt
upvotepartyparrot : Michael Schmidt
congapartyparrot : Michael Schmidt
openwhisk : Michael Schmidt
whisking : Michael Schmidt
yay : Michael Schmidt
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563984632140300?thread_ts=1563984632.140300&cid=C3TPCAQG1
----
2019-07-24 16:10:40 UTC - Matt Rutkowski:
<https://www.cbronline.com/news/apache-openwhisk-ibm>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563984640140600
----
2019-07-24 16:19:14 UTC - Dragos Dascalita Haut: when the request comes it maps
GET `/hello/world` to the `hello` action by using the data stored in Redis.
<https://github.com/apache/incubator-openwhisk-apigateway/blob/master/scripts/lua/routing.lua>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985154140800?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:20:57 UTC - chetanm: So upon each request it checks from redis
and then handles/adapts the request as per swagger defn stored in redis
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985257141000?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:21:36 UTC - Dragos Dascalita Haut: yes, but on the swagger part,
I don't think that's stored in Redis.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985296141300?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:22:08 UTC - chetanm: so in real prod is redis to be backed by
some persistent store like mysql
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985328141500?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:22:19 UTC - Dragos Dascalita Haut: no
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985339141700?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:22:20 UTC - chetanm: as all data in redis is ephermal
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985340141900?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:22:55 UTC - chetanm: so how does the data is retained post restart
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985375142100?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:23:08 UTC - Dragos Dascalita Haut: post Redis restart ?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985388142300?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:23:22 UTC - chetanm: yup
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985402142500?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:23:24 UTC - Dragos Dascalita Haut: well, for prod scenario Redis
should be clustered
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985404142700?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:24:27 UTC - chetanm: okie so we treat redis as a long term
storage here.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985467142900?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:24:35 UTC - Dragos Dascalita Haut: correct
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985475143100?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:25:11 UTC - chetanm: and is we run multiple separate clusters
then we need to think of some other approach
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985511143300?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:25:28 UTC - chetanm: like you had with minio based deployment to
use some object storage?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985528143500?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:25:35 UTC - Dragos Dascalita Haut: our team has an alternate impl
which uses static NGINX configuration files which are generated from the
swagger file provided by developers
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985535143700?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:25:43 UTC - chetanm: (minio as in dev tools .)
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985543143900?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:26:20 UTC - chetanm: so if I need to just have minimal api
gateway support in standalone
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985580144100?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:26:42 UTC - chetanm: I just need to start the apigw docker, redis
and configure the edge urls properly
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985602144300?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:26:52 UTC - chetanm: would nginx be needed?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985612144500?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:27:47 UTC - chetanm: i am missing the config where api gw docker
routes to controller
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563985667144700?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:42:16 UTC - Dragos Dascalita Haut: looking
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986536144900?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:42:39 UTC - Dragos Dascalita Haut: @chetanm are you looking to
start this docker from the _launchpad_ ?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986559145100?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:44:02 UTC - chetanm: From with in stand alone jar
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986642145300?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:45:37 UTC - Dragos Dascalita Haut: this is the config file that
has the managed APIs
<https://github.com/apache/incubator-openwhisk-apigateway/blob/master/conf.d/managed_endpoints.conf#L57-L95>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986737145500?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:46:40 UTC - Dragos Dascalita Haut: it's not the one that routes
to controller. I'm looking for that one
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986800145700?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:47:04 UTC - chetanm: Oh so it's hardwired to https 443
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986824145900?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:47:22 UTC - Dragos Dascalita Haut: there's one in ansible:
<https://github.com/apache/incubator-openwhisk/blob/master/ansible/roles/nginx/templates/nginx.conf.j2>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986842146100?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:48:01 UTC - Dragos Dascalita Haut: there's also one in
docker-compose:
<https://github.com/apache/incubator-openwhisk-devtools/blob/master/docker-compose/apigateway/generated-conf.d/whisk-docker-compose.conf>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986881146300?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:49:21 UTC - chetanm: Okie
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986961146500?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:49:26 UTC - Dragos Dascalita Haut: BTW, @chetanm I could also
help you with a minimalist set of configs to get the API GW up and running. you
can worry for the docker run command from the standalone jar, and I can help
with the configs
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986966146700?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:49:39 UTC - chetanm: Cool
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563986979146900?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:50:05 UTC - chetanm: Let me come up with base pr
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987005147100?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:50:12 UTC - Dragos Dascalita Haut: we'd also have to include in
that jar a set of resources like some nginx config files (target would be 1
config file to keep it simple )
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987012147300?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:50:33 UTC - chetanm: Okie
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987033147500?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:50:38 UTC - Dragos Dascalita Haut: I would also guess that we can
skip SSL/ `443` and go with another port ?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987038147700?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:50:44 UTC - chetanm: Yes
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987044147900?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:50:47 UTC - chetanm: No ssl
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987047148100?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:51:08 UTC - chetanm: 3233 is default port
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987068148300?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:52:28 UTC - Dragos Dascalita Haut: if that's the case then in
src/main/resources you can have a simple config file :
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987148148500?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:52:35 UTC - Dragos Dascalita Haut: ```upstream whisk_controller {
server whisk.controller:8888;
keepalive 16;
}
server {
listen 3233 default;
proxy_http_version 1.1;
proxy_set_header Connection "";
client_max_body_size 50m; # allow bigger functions to be deployed
location /docs {
proxy_pass <http://whisk_controller>;
}
location /api-docs {
proxy_pass <http://whisk_controller>;
}
location /api/v1 {
proxy_pass <http://whisk_controller>;
proxy_read_timeout 70s; # 60+10 additional seconds to allow controller to
terminate request
}
}```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987155148700?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:53:01 UTC - Dragos Dascalita Haut: replace
`whisk.controller:8888` with something else to reach the controller
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987181148900?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:54:49 UTC - chetanm: Okie
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987289149100?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:54:58 UTC - Dragos Dascalita Haut: say you have a file in
`src/main/resources/openwhisk-nginx.conf` then when executing `docker run` from
the jar mount this file like this: `-v
openwhisk-nginx.conf:/etc/api-gateway/conf.d/openwhis-nginx.conf`
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987298149300?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:55:30 UTC - Dragos Dascalita Haut: it should be automatically
included by NGINX. see
<https://github.com/apache/incubator-openwhisk-apigateway/blob/master/api-gateway.conf#L72>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987330149600?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:56:57 UTC - chetanm: So no explicit handling of
'^/api/([a-zA-Z0-9\-]+)/([a-zA-Z0-9'
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987417149800?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:57:21 UTC - Dragos Dascalita Haut: we/I can work on that once
this first part works
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987441150000?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:58:11 UTC - chetanm: Okie
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987491150200?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:58:13 UTC - Dragos Dascalita Haut: I'd assume that the definition
for that location would be automatically included, as the files are in `conf.d`
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987493150400?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 16:58:17 UTC - Dragos Dascalita Haut:
<https://github.com/apache/incubator-openwhisk-apigateway/tree/master/conf.d>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987497150600?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 17:03:14 UTC - chetanm: Okie
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987794150800?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 17:03:42 UTC - chetanm: What about the url which are as per some
swagger so not /api/v1
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987822151000?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 17:04:09 UTC - chetanm: Not sure if I know that part ...let me play
with that
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987849151200?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 17:04:43 UTC - chetanm: I am bit confused on api in which gateway
actually manipulates the incoming request and convert it to some URL for
controller
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563987883151400?thread_ts=1563979121.139300&cid=C3TPCAQG1
----
2019-07-24 18:33:55 UTC - Michael Schmidt: woot
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1563993235152400?thread_ts=1563984632.140300&cid=C3TPCAQG1
----
2019-07-24 20:26:45 UTC - Sam Hjelmfelt: This may help with Kubernetes
deployments. The container per second throughput of the YuniKorn scheduler is
much higher that other options. Batch processing requires large numbers of
“ephemeral” containers.
<https://blog.cloudera.com/blog/2019/07/yunikorn-a-universal-resource-scheduler/>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1564000005153900
----
2019-07-24 20:27:08 UTC - Sam Hjelmfelt:
<https://github.com/cloudera/yunikorn-core>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1564000028154200
----
