You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@trafficserver.apache.org by "Jon Cowie (JIRA)" <ji...@apache.org> on 2012/08/23 18:15:42 UTC
[jira] [Created] (TS-1416) Heartbeat increases Cache Misses count
Jon Cowie created TS-1416:
-----------------------------
Summary: Heartbeat increases Cache Misses count
Key: TS-1416
URL: https://issues.apache.org/jira/browse/TS-1416
Project: Traffic Server
Issue Type: Bug
Components: Cache
Affects Versions: 3.2.0
Environment: Centos 6.3
Reporter: Jon Cowie
Priority: Minor
With no network traffic going to the TS box, the following increase in Read Misses can be seen:
{code}
trafficserver> show:cache-stats
Bytes Used --- 0 GB
Cache Size --- 9 GB
--RAM Cache--
Total Bytes -- 10737418240
Bytes Used --- 0
Hits --------- 0
Misses ------- 0
--Lookups--
In Progress -- 0
Hits --------- 0
Misses ------- 0
--Reads--
In Progress -- 0
Hits --------- 0
Misses ------- 35
--Writes--
In Progress -- 0
Hits --------- 0
Misses ------- 0
--Updates--
In Progress -- 0
Hits --------- 0
Misses ------- 0
--Removes--
In Progress -- 0
Hits --------- 0
Misses ------- 0
{code}
After discussion with PSUdaemon, he identified this as occurring as follows:
{code}
<PSUdaemon> in iocore/cache/CacheRead.cc
<PSUdaemon> CACHE_INCREMENT_DYN_STAT(cache_read_failure_stat);
{code}
This isn't a particularly serious bug, but it does mean that when testing the box the Cache Miss count is artificially high.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (TS-1416) Heartbeat increases Cache Misses count
Posted by "Phil Sorber (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/TS-1416?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13440420#comment-13440420 ]
Phil Sorber commented on TS-1416:
---------------------------------
I was helping Jon look into this. The traffic_cop heartbeat is what is triggering these lookups. There is a config (proxy.config.http.record_heartbeat) that can be set to 0 to disable some HttpSM stats recording, but the stats he is pulling are from the cache module.
To fix this we could add another parameter to the cache lookup that dictates whether or not to maintain stats for a particular request. We could also modify the traffic_cop to be able to disable the heartbeat feature via a config variable. Just throwing out ideas here, amc may have a more elegant way to address this.
> Heartbeat increases Cache Misses count
> --------------------------------------
>
> Key: TS-1416
> URL: https://issues.apache.org/jira/browse/TS-1416
> Project: Traffic Server
> Issue Type: Bug
> Components: Cache
> Affects Versions: 3.2.0
> Environment: Centos 6.3
> Reporter: Jon Cowie
> Priority: Minor
>
> With no network traffic going to the TS box, the following increase in Read Misses can be seen:
> {code}
> trafficserver> show:cache-stats
> Bytes Used --- 0 GB
> Cache Size --- 9 GB
> --RAM Cache--
> Total Bytes -- 10737418240
> Bytes Used --- 0
> Hits --------- 0
> Misses ------- 0
> --Lookups--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Reads--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 35
> --Writes--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Updates--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Removes--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> {code}
> After discussion with PSUdaemon, he identified this as occurring as follows:
> {code}
> <PSUdaemon> in iocore/cache/CacheRead.cc
> <PSUdaemon> CACHE_INCREMENT_DYN_STAT(cache_read_failure_stat);
> {code}
> This isn't a particularly serious bug, but it does mean that when testing the box the Cache Miss count is artificially high.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (TS-1416) Heartbeat increases Cache Misses count
Posted by "Leif Hedstrom (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/TS-1416?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13440470#comment-13440470 ]
Leif Hedstrom commented on TS-1416:
-----------------------------------
In the grand scheme of things, these cache misses are noise :). But agreed, it'd be nice to exclude them from all stats, via that records_heartbeat setting. Disabling traffic_cop's healthcheck seems a bit odd, it's the whole point of running traffic_cop in the first place. If you don't want that feature, simply start ATS using traffic_manager directly (i.e. don't run traffic_cop at all).
> Heartbeat increases Cache Misses count
> --------------------------------------
>
> Key: TS-1416
> URL: https://issues.apache.org/jira/browse/TS-1416
> Project: Traffic Server
> Issue Type: Bug
> Components: Cache
> Affects Versions: 3.2.0
> Environment: Centos 6.3
> Reporter: Jon Cowie
> Priority: Minor
>
> With no network traffic going to the TS box, the following increase in Read Misses can be seen:
> {code}
> trafficserver> show:cache-stats
> Bytes Used --- 0 GB
> Cache Size --- 9 GB
> --RAM Cache--
> Total Bytes -- 10737418240
> Bytes Used --- 0
> Hits --------- 0
> Misses ------- 0
> --Lookups--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Reads--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 35
> --Writes--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Updates--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Removes--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> {code}
> After discussion with PSUdaemon, he identified this as occurring as follows:
> {code}
> <PSUdaemon> in iocore/cache/CacheRead.cc
> <PSUdaemon> CACHE_INCREMENT_DYN_STAT(cache_read_failure_stat);
> {code}
> This isn't a particularly serious bug, but it does mean that when testing the box the Cache Miss count is artificially high.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (TS-1416) Heartbeat increases Cache Misses count
Posted by "Leif Hedstrom (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/TS-1416?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Leif Hedstrom updated TS-1416:
------------------------------
Fix Version/s: 3.3.1
Why has this not been committed?
> Heartbeat increases Cache Misses count
> --------------------------------------
>
> Key: TS-1416
> URL: https://issues.apache.org/jira/browse/TS-1416
> Project: Traffic Server
> Issue Type: Bug
> Components: Cache
> Affects Versions: 3.2.0
> Environment: Centos 6.3
> Reporter: Jon Cowie
> Priority: Minor
> Fix For: 3.3.1
>
> Attachments: cop_no_cache_lookup.diff
>
>
> With no network traffic going to the TS box, the following increase in Read Misses can be seen:
> {code}
> trafficserver> show:cache-stats
> Bytes Used --- 0 GB
> Cache Size --- 9 GB
> --RAM Cache--
> Total Bytes -- 10737418240
> Bytes Used --- 0
> Hits --------- 0
> Misses ------- 0
> --Lookups--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Reads--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 35
> --Writes--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Updates--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Removes--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> {code}
> After discussion with PSUdaemon, he identified this as occurring as follows:
> {code}
> <PSUdaemon> in iocore/cache/CacheRead.cc
> <PSUdaemon> CACHE_INCREMENT_DYN_STAT(cache_read_failure_stat);
> {code}
> This isn't a particularly serious bug, but it does mean that when testing the box the Cache Miss count is artificially high.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (TS-1416) Heartbeat increases Cache Misses count
Posted by "Phil Sorber (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/TS-1416?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13450172#comment-13450172 ]
Phil Sorber commented on TS-1416:
---------------------------------
02024e4 TS-1416: Do not do cache lookup if cop_test_page is true to stop artificial increase of cache misses.
> Heartbeat increases Cache Misses count
> --------------------------------------
>
> Key: TS-1416
> URL: https://issues.apache.org/jira/browse/TS-1416
> Project: Traffic Server
> Issue Type: Bug
> Components: Cache
> Affects Versions: 3.2.0
> Environment: Centos 6.3
> Reporter: Jon Cowie
> Priority: Minor
> Fix For: 3.3.1
>
> Attachments: cop_no_cache_lookup.diff
>
>
> With no network traffic going to the TS box, the following increase in Read Misses can be seen:
> {code}
> trafficserver> show:cache-stats
> Bytes Used --- 0 GB
> Cache Size --- 9 GB
> --RAM Cache--
> Total Bytes -- 10737418240
> Bytes Used --- 0
> Hits --------- 0
> Misses ------- 0
> --Lookups--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Reads--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 35
> --Writes--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Updates--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> --Removes--
> In Progress -- 0
> Hits --------- 0
> Misses ------- 0
> {code}
> After discussion with PSUdaemon, he identified this as occurring as follows:
> {code}
> <PSUdaemon> in iocore/cache/CacheRead.cc
> <PSUdaemon> CACHE_INCREMENT_DYN_STAT(cache_read_failure_stat);
> {code}
> This isn't a particularly serious bug, but it does mean that when testing the box the Cache Miss count is artificially high.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira