[icecast-dev] Icecast's YP bugs

Ricardo Galli gallir at uib.es
Tue Jun 24 17:08:08 PDT 2003

On Wednesday 25 June 2003 01:58, Arc shaped the electrons to say:
> Ok I've been discovering many bugs in Icecast's YP implementation..
> The most serious of which is, if a YP server is rejecting a stream the
> server has a memory leak which, as it continues to re-try to publish
> it's rejected streams, grows until the server segfaults.  On our server,
> for instance, we're relaying shoutcast streams (which does not relay the
> stream name and other metadata) which are rejected from the YP server
> despite that they are comming from an Icecast2 server and using proper
> protocols because of the missing data.  This grows over time and, as a
> result, our relay server continues to segfault.
> Also, last night I was working on the YP server just before our server
> suffered from some network issue and was taken offline (but still
> working).  A stream source running from the box (from a playlist)
> continued to run to our main Icecast process which kept getting rejected
> from the YP server.. when we got the system back online I found this:
> 37088 nobody     2   0   213M 47696K RUN     55:12  0.00%  0.00% icecast
> Here are the bugs as I've found so far...
> 1) Icecast memory leak, specifically caused by rejected YP add/touches

I've sent a patch ages ago (around september-october last year, check 
archives) that solves one side of this proble. It's a one-liner.

Easealy reproductable: do a script that try to connect to a non-existing 
stream or that connects and disconnect inmediately. In few ours you will see 
the OOM killer. 

  ricardo galli       GPG id C8114D34
--- >8 ----
List archives:  http://www.xiph.org/archives/
icecast project homepage: http://www.icecast.org/
To unsubscribe from this list, send a message to 'icecast-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.

More information about the Icecast-dev mailing list