Re: Thread affinity on Continuations vs TSContSchedule API

2020-03-27 Thread Alan Carroll
I put most of my discussion on the PR 6577, but discussing this with Sudheer and Fei, I think the consensus is 1) Accept the PR for ATS 9. 2) For ATS 10 (or maybe 9.1?), deprecate TSContSchedule. Make it explicit to schedule on a pool or thread. On Thu, Mar 26, 2020 at 2:51 PM Walt Karas wrote:

Re: Thread affinity on Continuations vs TSContSchedule API

2020-03-27 Thread David Calavera
Something I find challenging about TSContScheduleOnPool and others is that there is no much documentation about good practices to configure thread pools. For example the only documentation I found about task threads only says that you must have at least 1 task thread: https://docs.trafficserver.ap

Re: Thread affinity on Continuations vs TSContSchedule API

2020-03-27 Thread Alan Carroll
Yes, that's a problem. One of the many things on my list. On Fri, Mar 27, 2020 at 12:42 PM David Calavera wrote: > Something I find challenging about TSContScheduleOnPool and others is that > there is no much documentation about good practices to configure thread > pools. For example the only do

Re: Thread affinity on Continuations vs TSContSchedule API

2020-03-27 Thread Sudheer Vinukonda
Generally speaking, we schedule Continuations for 2 main use cases.  1) Make async sideways calls 2) Some offloading of heavy CPU work (such as encrypting/decrypting a token etc) For (1) we don't use separate thread pools and just schedule them on the Net threads. As these involve i/o, it's more

Re: Proposal for utility to avoid use of stale pointers to objects that have been destroyed.

2020-03-27 Thread Walt Karas
Here is some (simplistic) example code using NoStale: https://github.com/ywkaras/MiscRepo/blob/master/NoStale/example.cc On Thu, Mar 26, 2020 at 11:53 AM Walt Karas wrote: > https://github.com/ywkaras/MiscRepo/blob/master/NoStale/NoStale.h >