You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openwhisk.apache.org by OpenWhisk Team Slack <ra...@apache.org> on 2019/11/28 09:29:40 UTC
[slack-digest] [2019-11-27] #general
2019-11-27 01:35:39 UTC - Dominic Kim: Is there any material that I can F/U about the box and logs things?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574818539465800
----
2019-11-27 02:02:43 UTC - Rodric Rabbah: Thanks Dominic. It seems to be working. The caveat is the folder needs period purging. There’s no way to configure that from the box app at least without upgrading. So it’s manual.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574820163467400
----
2019-11-27 02:03:23 UTC - Dominic Kim: Can I have the access as well?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574820203467800
----
2019-11-27 02:20:15 UTC - Rodric Rabbah: Yep will add you (dm me your preferred email)
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574821215468200
----
2019-11-27 12:50:21 UTC - Dave Grove: The load balancer tries to maximize locality by assigning multiple invocations of a function to the same node. It only will use a second invoker when the first invoker is “full”. I think you should configure your cluster so that each invoker has 16gb of memory for user actions and increase number of invokers to match overall cluster capacity. When that is working, fine tune by decreasing number of invoker and increasing memory per invoker until it starts to degrade
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574859021476300?thread_ts=1574633042.430600&cid=C3TPCAQG1
----
2019-11-27 15:04:30 UTC - Bill Zong: No tech interchange meeting today?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574867070477500
----
2019-11-27 15:07:12 UTC - Bill Zong: Everybody in Thanksgiving day holiday?:joy:
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574867232478100
----
2019-11-27 15:13:20 UTC - Rodric Rabbah: right - none today, next week
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574867600478400
----
2019-11-27 15:13:42 UTC - Rodric Rabbah: (deleted it :slightly_smiling_face:)
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574867622478700
----
2019-11-27 15:36:34 UTC - Denis Kokorin: OpenWhisk documentation stays that serverless architecture fits microservices. But microservices have to store data somewhere. Because container can be stopped at any time I can't use connection pooling or something similar
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574868994481300
----
2019-11-27 15:36:58 UTC - Denis Kokorin: Could you recommend any readings on that topic?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574869018481800?thread_ts=1574869018.481800&cid=C3TPCAQG1
----
2019-11-27 16:01:20 UTC - Ali Tariq: finally some progress! i decreased the containerpool to 220 with 5 invoker replicas (initially i did not know that the containerpool value is specific to per invoker which was why all 5 replicas were being initialized with 1100 containers capacity. This is precisely why invoker-4 was being overwhelm with load). Now invoker never goes down for maintainence but still only a single invoker replica is being used which is limiting the concurrency to 220 containers. (invoker-0 being full does not result in the use of other invokers, rather is waits for invoker-0 to get free again)
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574870480481900?thread_ts=1574633042.430600&cid=C3TPCAQG1
----
2019-11-27 20:19:16 UTC - Upkar Lidder: I don’t have any readings, but the idea is for the functions to be stateless. This does not mean that they cannot read / write data. You could have a function/microservice that reads data from a database and another function/microservice that writes data to a database (like nosql or rdbms). You can also read data from write data to a file storage (like s3 or IBM COS). Another architecture would be to use some of the pre-built pakages in openWhisk like openwhisk-package-cloudant. Lastly, you can use a message hub like Kafka to loosely couple your different microservices so that a function could process data and another would interface with the storage solution once the processing is done.
+1 : Dominic Kim
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574885956482500?thread_ts=1574869018.481800&cid=C3TPCAQG1
----
2019-11-27 20:25:52 UTC - Upkar Lidder: When you say `container can be stopped at any time`, there is an upper limit in OpenWhisk of how long a function can run. This is listed in the docs as 5 mins (<https://github.com/apache/openwhisk/blob/master/docs/reference.md#system-limits>). The hosted version on IBM Cloud increases this to 10 mins (<https://cloud.ibm.com/docs/openwhisk?topic=cloud-functions-limits>). Of course a function may stop if it fails for some other reason.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574886352482700?thread_ts=1574869018.481800&cid=C3TPCAQG1
----
2019-11-27 23:57:03 UTC - Ali Tariq: could it be because i am using `whisk:ingress:type:NodePort` instead of `whisk:ingress:type:Loadbalancer` ? if so, how do i configure this change? i tried changing the `type` value but that gives me `getsockopt: connection refused` for every `wsk` command.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1574899023482900?thread_ts=1574633042.430600&cid=C3TPCAQG1
----