Re: TS-4703 && PR-829

2016-09-09 Thread James Peach
> On Sep 9, 2016, at 2:22 PM, Alan Carroll > wrote: > > 1. So why the output array then? > > >On Friday, September 9, 2016 1:19 PM, James Peach > wrote: > > > >> On Sep 9, 2016, at 10:55 AM, Alan Carroll >> wrote: >> >> No. I'm not completely sure what I meant when I wrote that

Re: TS-4703 && PR-829

2016-09-09 Thread Alan Carroll
1. On Friday, September 9, 2016 1:19 PM, James Peach wrote: > On Sep 9, 2016, at 10:55 AM, Alan Carroll > wrote: > > No. I'm not completely sure what I meant when I wrote that (maybe it was a > typo). My view on review was that it should be an in/out parameter where the > tag (poss

Re: performance problem with traffic server 6.2

2016-09-09 Thread Alan Carroll
Based on this data the most likely explanation is a plugin that didn't call TSHttpTxnReenable(). On Friday, September 9, 2016 2:04 PM, Shu Kit Chan wrote: Is there any plugins you are using? On Fri, Sep 9, 2016 at 10:10 AM, Siddharth Agarwal wrote: > Hi, > > We are on the traffic se

Re: performance problem with traffic server 6.2

2016-09-09 Thread Shu Kit Chan
Is there any plugins you are using? On Fri, Sep 9, 2016 at 10:10 AM, Siddharth Agarwal wrote: > Hi, > > We are on the traffic server 6.2 release. We are experiencing performance > issues with this release and have enabled slow request logging. Our health > check end point sometimes is taking taki

Re: TS-4703 && PR-829

2016-09-09 Thread James Peach
> On Sep 9, 2016, at 10:55 AM, Alan Carroll > wrote: > > No. I'm not completely sure what I meant when I wrote that (maybe it was a > typo). My view on review was that it should be an in/out parameter where the > tag (possibly just a prefix) is passed in and the actual tag found is passed >

Re: TS-4703 && PR-829

2016-09-09 Thread Alan Carroll
No. I'm not completely sure what I meant when I wrote that (maybe it was a typo). My view on review was that it should be an in/out parameter where the tag (possibly just a prefix) is passed in and the actual tag found is passed back. Susan has convinced me it would be better to make split that

Re: TS-4703 && PR-829

2016-09-09 Thread James Peach
> On Sep 9, 2016, at 10:15 AM, Susan Hinrichs > wrote: > > > > On 9/9/2016 12:01 PM, James Peach wrote: >>> On Sep 9, 2016, at 9:39 AM, Susan Hinrichs >>> wrote: >>> >>> I'd like to propose a slight tweak for TSHttpTxnClientProtocolStackContains >>> and TSHttpSsnClientProtocolStackContain

Re: TS-4703 && PR-829

2016-09-09 Thread Susan Hinrichs
On 9/9/2016 12:01 PM, James Peach wrote: On Sep 9, 2016, at 9:39 AM, Susan Hinrichs wrote: I'd like to propose a slight tweak for TSHttpTxnClientProtocolStackContains and TSHttpSsnClientProtocolStackContains Replace the tag argument with two explicit input and output arguments, e.g. int T

performance problem with traffic server 6.2

2016-09-09 Thread Siddharth Agarwal
Hi, We are on the traffic server 6.2 release. We are experiencing performance issues with this release and have enabled slow request logging. Our health check end point sometimes is taking taking more than 3 seconds to respond. Our CPU is idle. We are seeing most of time spent in plugin_total whil

Re: TS-4703 && PR-829

2016-09-09 Thread James Peach
> On Sep 9, 2016, at 9:39 AM, Susan Hinrichs > wrote: > > I'd like to propose a slight tweak for TSHttpTxnClientProtocolStackContains > and TSHttpSsnClientProtocolStackContains > > Replace the tag argument with two explicit input and output arguments, e.g. > > int TSHttpTxnClientProtocolStac

Re: TS-4703 && PR-829

2016-09-09 Thread Susan Hinrichs
I'd like to propose a slight tweak for TSHttpTxnClientProtocolStackContains and TSHttpSsnClientProtocolStackContains Replace the tag argument with two explicit input and output arguments, e.g. int TSHttpTxnClientProtocolStackContains(TSHttpTxn txnp, char const *contains_tag, char const **spec