[
https://issues.apache.org/jira/browse/TS-3413?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Susan Hinrichs updated TS-3413:
-------------------------------
Attachment: ts-3413.diff
Here is the fix that adjusts how HTTP/2 users FetchSM to handle the EOS case.
It solves the problem for me.
[~zwoop] was running this fix along with the memory fixes from TS-3405 on
doc.trafficserver. However, he was not reporting the CPU utilization problem,
so the best I can say from that is that this change does not introduce
instability.
[~rokubo] and [~briang] can you take a look at the patch ts-3413.diff and let
me know if it seems reasonable?
> High CPU utilization when processing Http/2 traffic
> ---------------------------------------------------
>
> Key: TS-3413
> URL: https://issues.apache.org/jira/browse/TS-3413
> Project: Traffic Server
> Issue Type: Bug
> Components: HTTP/2, Performance
> Reporter: Susan Hinrichs
> Attachments: ts-3413.diff
>
>
> Running ATS in forward transparent mode with Http/2 support enabled. After a
> few page views, ATS gets into a state where it is consistently using more
> than 100% CPU. I think what is happening is that the origin server is closing
> the connection (we get a TS_EVENT_VCONN_EOS in FetchSM). FetchSM sends a
> BODY_DONE to Http2ClientSession but the http_vc (origin socket?) is never
> closed, so with each poll interval the EOS is reported again and again.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
