Yes, I’d highly recommend against using this plugin. Not only is it broken in many ways,
it’s no longer supported in new versions of ATS. And, it really doesn’t do you much good
anyways, I’d look at the collapsed forwarding plugin(s) instead.
Cheers,
— Leif
> On Nov 17, 2018, at 1:01 AM, Alan Carroll <solidwallofcode@oath.com> wrote:
>
> Looks like a problem with the "stale-while-revalidate" plugin. I know that's been problematic.
I'm not sure what the current state is.
>
>> On Fri, Nov 16, 2018 at 8:43 AM Vasanth Mathivanan <vasanth.m@evolutiondigital.com>
wrote:
>> Yes , its generating core dump file in traffic server root path and we are using
ATS.6.2 in centos 6.5.
>>
>> traffic.out logs
>>
>> “ATAL: InkAPI.cc:1879: failed assert `sdk_sanity_check_mbuffer(bufp) == TS_SUCCESS`
>>
>> traffic_server: using root directory '/opt/trafficserver'
>>
>> traffic_server: received signal 6 (Aborted)
>>
>> traffic_server - STACK TRACE:
>>
>> /opt/trafficserver/bin/traffic_server(_Z19crash_logger_invokeiP7siginfoPv+0x8e)[0x4ac5ee]
>>
>> /lib64/libpthread.so.0[0x3030e0f7e0]
>>
>> /lib64/libc.so.6(gsignal+0x35)[0x3030a32495]
>>
>> /lib64/libc.so.6(abort+0x175)[0x3030a33c75]
>>
>> /opt/trafficserver/lib/libtsutil.so.6(+0x28228)[0x2b5cb60dc228]
>>
>> /opt/trafficserver/lib/libtsutil.so.6(+0x282bc)[0x2b5cb60dc2bc]
>>
>> /opt/trafficserver/lib/libtsutil.so.6(+0x26595)[0x2b5cb60da595]
>>
>> /opt/trafficserver/bin/traffic_server(TSHandleMLocRelease+0x3a)[0x4c33ca]
>>
>> /opt/trafficserver/libexec/trafficserver/stale_while_revalidate.so(+0x38c9)[0x2b5e9eaa48c9]
>>
>> /opt/trafficserver/bin/traffic_server(_ZN15INKContInternal12handle_eventEiPv+0xdb)[0x4c12ab]
>>
>> /opt/trafficserver/bin/traffic_server(_ZN6HttpSM17state_api_calloutEiPv+0x1b7)[0x5ac967]
>>
>> /opt/trafficserver/bin/traffic_server(_ZN6HttpSM14set_next_stateEv+0x65b)[0x5b226b]
>>
>> /opt/trafficserver/bin/traffic_server(_ZN6HttpSM22state_cache_open_writeEiPv+0x20e)[0x5a3cde]
>>
>> /opt/trafficserver/bin/traffic_server(_ZN6HttpSM12main_handlerEiPv+0xca)[0x5b3f3a]
>>
>> /opt/trafficserver/bin/traffic_server(_ZN11HttpCacheSM22state_cache_open_writeEiPv+0x281)[0x58a481]
>>
>> /opt/trafficserver/bin/traffic_server(_ZN7EThread13process_eventEP5Eventi+0x120)[0x766f00]
>>
>> /opt/trafficserver/bin/traffic_server(_ZN7EThread7executeEv+0x811)[0x767bd1]
>>
>> /opt/trafficserver/bin/traffic_server[0x76698a]
>>
>> /lib64/libpthread.so.0[0x3030e07aa1]
>>
>> /lib64/libc.so.6(clone+0x6d)[0x3030ae8bdd]”
>>
>>
>>
>> No I did not set ephemeral ports .
>>
>>
>> Had default values already .
>>
>> net.ipv4.ip_local_port_range = 32768 60999
>>
>> net.ipv4.ip_local_reserved_ports = ###this one is empty
>>
>>
>>
>> So its set by manually .
>>
>> net.ipv4.ip_local_reserved_ports = 30000-31000
>>
>>
>>
>> Is this ok ? .
>>
>>
>>
>>
>> __Vasanth
>>
>> Sent from Mail for Windows 10
>>
>>
>>
>> From: Susan Hinrichs
>> Sent: Friday, November 16, 2018 7:55 PM
>> To: users@trafficserver.apache.org
>> Subject: Re: Traffic server suddenly drop the packets and crashed
>>
>>
>>
>> Are you exhausting ephemeral ports? What is the output of "ss -s" on your Traffic
Server machine? That is quite easy exhaust all of the ephemeral ports for a single load testing
client machine between active and time-wait connections.
>>
>> Is Traffic Server generating a core or a stack trace? What version of Traffic Server?
>>
>>> On Fri, Nov 16, 2018 at 1:42 AM Vasanth Mathivanan <vasanth.m@evolutiondigital.com>
wrote:
>>> Hi ,
>>>
>>> We are testing 1000 client access one by one load test after 700 clients dropping
the packets like “connection timeout” or “Connection reset by peer ” errors .
>>>
>>>
>>>
>>> Ram and Cpus resources are everything is fine .Ethernet speed also 20gb/s both
client and traffic server . Origin server has 2gb/s speed between ATS and Origin server
.While running 700 requests in origin server content will be receiving speed less than 10Mb/s
then traffic server and client getting more than 1Gb/s speed .
>>>
>>>
>>>
>>> We are also monitoring that resources everything running fine but crashed the
traffic server after 700 requests .
>>>
>>>
>>>
>>> Note: We are created Java player , We will give the no of requests its call
the particular URL based on given requests .
>>>
>>>
>>>
>>> __Vasanth
>>>
>>>
>>>
>>> Sent from Mail for Windows 10
>>>
>>>
>>>
>
>
> --
> Beware the fisherman who's casting out his line in to a dried up riverbed.
> Oh don't try to tell him 'cause he won't believe. Throw some bread to the ducks instead.
> It's easier that way. - Genesis : Duke : VI 25-28
|