You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@lucene.apache.org by "Ishan Chattopadhyaya (Jira)" <ji...@apache.org> on 2019/12/11 05:20:00 UTC
[jira] [Comment Edited] (SOLR-14050) clean up tests use of network
addresses
[ https://issues.apache.org/jira/browse/SOLR-14050?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16993197#comment-16993197 ]
Ishan Chattopadhyaya edited comment on SOLR-14050 at 12/11/19 5:19 AM:
-----------------------------------------------------------------------
Oh yes, this patch made the tests fast. Took 5m28s (as opposed to 7m+) usually! This was with security manager turned off.
Btw, FYI, when security manager was turned on, the tests showed many:
{code}
[junit4] 2> java.security.AccessControlException: access denied ("java.io.FilePermission" "/usr/share/java/ant.jar" "read")
{code}
Edit: And second run was 5m09s. This patch is awesome, +1 to commit!
was (Author: ichattopadhyaya):
Oh yes, this patch made the tests fast. Took 5m28s (as opposed to 7m+) usually! This was with security manager turned off.
Btw, FYI, when security manager was turned on, the tests showed many:
{code}
[junit4] 2> java.security.AccessControlException: access denied ("java.io.FilePermission" "/usr/share/java/ant.jar" "read")
{code}
> clean up tests use of network addresses
> ---------------------------------------
>
> Key: SOLR-14050
> URL: https://issues.apache.org/jira/browse/SOLR-14050
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Robert Muir
> Priority: Major
> Attachments: SOLR-14050.patch
>
>
> Motivated by this tweet: https://twitter.com/ichattopadhyaya/status/1204274908454219778
> I think we should clean up the "connect/resolve anywhere" security permissions in the tests and make it stricter like lucene. This way we can prevent tests from doing slow things now or in the future.
> The biggest issue I see exploring this so far is solr has lots of tests that want to hit a "dead" node. If you get lucky test passes quickly. If you get unlucky (e.g. it tries to route that multicast ipv6 address somewhere?) then it passes slowly or maybe even times out or misbehaves.
> I think, depending on the networking environment (e.g. working ipv6 or not, maybe OS), the current way its being done can be terribly slow. I have some ideas, just need to do some testing.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org
For additional commands, e-mail: issues-help@lucene.apache.org