Eric - we also noticed that with OSSEC off we are still getting IO errors 
reported from Syslog-ng.

Still investigating..

D.

-----Original Message-----
From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL 
PROTECTED]
Sent: Tuesday, September 23, 2008 4:31 PM
To: ossec-list
Subject: [ossec-list] Re: "connection reset by peer(104)" condition


Trying to follow as I have the same error message from an agent that working 
fine for months.

Agents are running v1.4 so is server. Unfortnately I don't have a file
- /var/ossec/queue/shared.

Any ideas?
Thanks,
Eric



On Sep 18, 6:34 am, "Daniel Cid" <[EMAIL PROTECTED]> wrote:
> Hi Donald,
>
> Which version of OSSEC are you using (on the agent and server)? If
> thismessageis appearing at the sever side, can you show us the output
> of /var/ossec/queue/shared ? It means that the agent is sending an
> improper formatedmessage(not \nterminated), which might be caused by a
> truncated output (I only saw this in the past when the shared
> directory had a very large amount of files).
>
> *btw, how often are you seeing those?
> **is your first issue resolved?
>
> Thanks,
>
> --
> Daniel B. Cid
> dcid ( at ) ossec.net
>
> On Wed, Sep 17, 2008 at 4:13 AM, Donald Tabone <[EMAIL PROTECTED]> wrote:
>
> > Anyone seen this behavior before and managed to resolve it ?
>
> > On Sep 16, 10:54 am, "Donald Tabone" <[EMAIL PROTECTED]> wrote:
> >> Whilst we have suppressed the rule <1002> from sending email
> >> alerts, we also noticed the following log entry is
> >> /var/log/ossec.log
>
> >> 2008/09/16 01:47:50ossec-remoted:Invalidmessagefrom '10.31.19.33'
> >> (strchr\n)
>
> >> Has this got anything to do with syslog-ng/ossec generating errors.
>
> >> 2008/9/16 Donald Tabone <[EMAIL PROTECTED]>
>
> >> > Dear all,
>
> >> > We're experiencing this randomly generated error after we fire-up
> >> > ossec. It is tending to flood our email box with alerts.
>
> >> > Sep 15 10:03:32 captain syslog-ng[22504]: I/O error occurred
> >> > while reading; fd='69', error='Connection reset by peer (104)'
>
> >> > From some research we've done, using lsof the file descriptor
> >> > (fd) number corresponds at any one point in time to a number of
> >> > other files with different permissions. Through logical deduction
> >> > we think that it is possible that when ossec tries to check a
> >> > file on a particular channel, the file would have already been
> >> > closed and hence the connection is reset. Are off track with this line 
> >> > of thinking?
>
> >> > We also referenced
> >> >http://www.ossec.net/wiki/index.php/Know_How:Email_Alerts_below_7
> >> > and tried to tweak a rule as explained in the wiki entry - but
> >> >this  did not work. The result atm is that ossec service is stopped.
>
> >> > The questions we ask:
>
> >> > How can we accurately pinpoint whether ossec is the problem or not?
> >> > Is this a false positive? ie. can it be ignored or is there a
> >> > significant problem in the system we actually need to resolve?
>
> >> > We are running ossec 1.6 on a debian4 machine. (same error
> >> > occured with 1.5.1)
>
> >> > Can someone assist in finding a solution?
> >> > TY

Reply via email to