Subject: Re: SysMES detected BMC errors Content-Type: text/plain; charset="us-ascii"; format=flowed; delsp=yes In-Reply-To: <25B2CFAAEF514F4CB8B2481D800BE07D@firebat> Date: Thu, 23 Jun 2011 12:50:06 +0200 CC: "'alice-hlt-cluster-admin (ALICE HLT Cluster Admin List)'" Content-Transfer-Encoding: 7bit Message-ID: <9F692E51-FB3A-4784-858E-EA53A5F10F3F@cern.ch> References: <246279427.471308729668758.JavaMail.sysmes@mon1.internal> <40E73F0F-C9F7-459F-BA94-706675D67740@cern.ch> <25B2CFAAEF514F4CB8B2481D800BE07D@firebat> To: X-Mailer: Apple Mail (2.1076) MIME-Version: 1.0 On Thu, Jun 23, 2011, at 10:49:35AM GMT+02:00, Jochen Ulrich wrote: > Yes, it is a false-positive but it's not the SysMES monitor that > produces it but it's the BMC itself. > The SysMES monitor simply reads the BMC log and sends every > "critical" or "non-revocerable" message it finds in the log. Still -- couldn't it just poll the sensor directly when SysMES triggers on the "critical"- and "non-recoverable"-messages in the log? So something like this? Log entry triggers SysMES -> Wait 1 minute -> Poll sensors directly, check values -> Wait 1 minute -> Poll sensors directly, check values - > If still below threshold, send alarm -- Joachim