couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Marcello Barnaba (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (COUCHDB-1098) Crash when rapidly creating/populating/querying a database
Date Tue, 22 Mar 2011 16:49:05 GMT

    [ https://issues.apache.org/jira/browse/COUCHDB-1098?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13009711#comment-13009711
] 

Marcello Barnaba commented on COUCHDB-1098:
-------------------------------------------

Yeah - sorry I misunderstood the report from my colleague - tests on couchone were really
unlikely to fail due to the much higher latency between the client and the server.

So, this is definitely a race condition when the database is stressed very rapidly. It could
sound like a corner case, but it could happen on a busy application as well. Not WRT the DELETE/PUT
sequence, but WRT the POST/GET one - think about an HTTP server exposing a REST API, and that's
exactly what we're doing on our app. :-)

> Crash when rapidly creating/populating/querying a database
> ----------------------------------------------------------
>
>                 Key: COUCHDB-1098
>                 URL: https://issues.apache.org/jira/browse/COUCHDB-1098
>             Project: CouchDB
>          Issue Type: Bug
>          Components: Database Core
>    Affects Versions: 1.0.2
>         Environment: Server: CouchDB v1.0.2 on Erlang R14B on MacOS X 10.6.6 - via the
CouchDBX.app bundle v1.0.2.0
> Client:  Ruby (CouchRest)
>            Reporter: Marcello Barnaba
>            Priority: Critical
>
> Hello,
> in our test cases we repeatedly delete, recreate, populate and query the test database
views to assert the correctness of the data returned. For each test group.
> We've experiencing two, *random*, different crashes of the erlang processes, one during
indexing (I think) of the test data, and another one caused by a race condition when deleting
and recreating the database - when the test group ends. On this gist https://gist.github.com/ffc6960a2b7022826388
there's the relevant part of the couchdb access log file, from the beginning to the end of
the test group.
> Please let us know any other information that you need in order to help pinpointing the
issue.
> As a side note, it happens on another 1.0.2 installation on Linux/i386 and on couchone.com
servers - so it doesn't look like platform specific :-).
> Thank you in advance for your support.
> ~Marcello

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message