Hi All,

I have started to encounter a problem with a RELP based rsyslog server
which was described in a post to this list about 6 months ago. I have
similar releases of rsyslog mentioned in the post.

Before swinging into debug mode, I wanted to see if the problem has been
identified. So, has anyone encountered a similar problem? Was it
resolved by going to a more recent release?

The old post is below.

Any advise is welcome.

Rgds
Burn Alting

        [rsyslog] 5.8.10 RELP server exhausting connections and stops
        responding
        Julio Lajara ju2wheels at gmail.com 
        Wed Jun 13 19:00:10 CEST 2012 
        
        Previous message: [rsyslog] rsyslog 6.2.2 (v6-stable) released
              * Next message: [rsyslog] statsobj.c
              * Messages sorted by: [ date ] [ thread ] [ subject ]
        [ author ] 
              * 
        Hi all,
        
        Ive run into an issue with RELP where the server side doesnt
        close out 
        stale TCP connections (in particular its isolated to
        connections 
        initiated by clients that go through a firewall that according
        to my 
        network admins is configured to close TCP sessions that have
        been idle 
        for 30 min or more). We noticed that connections that didnt go
        through a 
        firewall dont seem to have this issue. During normal operation,
        our 
        server uses about 220 TCP connections and when it stops
        responding after 
        of few days of running it has around 800-1000, most of which are
        just 
        old idle connections that are never closed.
        
        I found a similar thread here where it happens on plain TCP
        transport: 
        http://lists.adiscon.net/pipermail/rsyslog/2010-July/011987.html
        
        1. If I use $ActionSendTCPRebindInterval, will this also apply
        to RELP 
        or does it only affect the normal TCP transport?
        2.  (Question from my network admin, excuse my wording if it
        sounds 
        weird:) Besides the above setting is there any option on the
        client side 
        to tell the client to reuse a previous existing connect that the
        server 
        side keeps open by sending a "RST" to the TCP connection if it
        already 
        isnt? I looked through the doc and the only option I found even
        close 
        was the one above, just want to make sure I didnt miss any
        other 
        possible options.
        
        If anyone has other suggestions on how to possibly resolve this,
        I would 
        be most grateful to hear about them.
        
        Thanks,
        

_______________________________________________
rsyslog mailing list
http://lists.adiscon.net/mailman/listinfo/rsyslog
http://www.rsyslog.com/professional-services/
What's up with rsyslog? Follow https://twitter.com/rgerhards
NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of 
sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T LIKE 
THAT.

Reply via email to