trafficserver-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From avinash katika <avinash.kat...@gmail.com>
Subject When DNS lookup (for Origin IP) is required and it takes time (Tx) to complete, a possible thunder herd for the same object which occurred during this DNS look up time (Tx) resulted in all cache misses to Origin.
Date Tue, 22 Jul 2014 23:36:13 GMT
Read the documentation about mitigating thunder herd problem and set the
parameters accordingly

Current Version : 4.2.1

CONFIG proxy.config.cache.enable_read_while_writer INT 1
CONFIG proxy.config.http.background_fill_active_timeout INT 0
CONFIG proxy.config.http.background_fill_completed_threshold FLOAT 0.000000
CONFIG proxy.config.cache.max_doc_size INT 0
----
CONFIG proxy.config.http.cache.max_open_read_retries INT 5
CONFIG proxy.config.http.cache.open_read_retry_time INT 10

We have  100 users making simultaneous requests(~1ms apart) for a single
object X (which is a TCP_MISS) to the ATS but Still see all the 100
requests going back to Origin with each request making a separate
connection to origin.Below is a more detailed explanation

*Test Setup:*

1. At time 10:00.000000, 1st request of object X arrives at ATS cache
2. 99 more requests for object X arrive between 10:00.*000* and
10:00.*100* (~1ms
apart)
3. ATS does origin DNS lookup at 10:00.*017000* and got the response at
10:01.*167000*(we purposely set up DNS look up for so long)
4. At 10:01.*167000*, 100 cache misses to Origin (for the same object X)
and each rides on its own TCP connection


Is this a know issue with ATS or are there any configuration parameters to
fix this issue?

Mime
View raw message