From: Edi CAHYADI (edi.cahyadi@st.com)
Date: Fri Mar 31 2006 - 04:56:38 EST
Hi list,
I received only one reply from Bob Vance. Thanks, Bob.
His email below explains it all.
With MAX_CONFIGURED_PACKAGES=5, I found a script run as a cron job had caused the messages, although actually I didn't find any harm to the cluster.
Regards,
Edi
-----Original Message-----
From: Bob Vance [mailto:bobvance@alumni.caltech.edu]
Sent: Thursday, March 30, 2006 10:21 PM
To: 'Edi CAHYADI'
Subject: RE: [HPADM] cmcld: Stopped accepting local connection requests
see thread http://forums2.itrc.hp.com/service/forums/questionanswer.do?threadId=103923
Currently, it takes 15 seconds for a connection to be released, so many at one time could reach the max. However, I notice that some connections are re-used in the new version, 11.15. Hmmm
Using 'lsof'
(you can download 'lsof' from the HP porting archive,
http://hpux.cs.utah.edu/
),
I see normally 5 connections to hacl-local.
'cmviewcl' causes one more.
15 extra seems a lot.
Maybe some EMS connections at one time?
Snippet from HP Knowledge Base document ================================================
Submitted Date: 4/6/00
Title: MC/SG: cmcld: stopped accepting local connection requests Document ID: S3100005230 Last Modified Date: 8/5/05
You may provide feedback on this document
PROBLEM
My ServiceGuard (SG) cluster generates the following message:
cmcld: Stopped accepting local connection requests because there
are currently too many concurrent connections to the
daemon (20).
Why am I running out of connections?
CONFIGURATION
Operating System - HP-UX
Version - 11.0
Hardware System - HP9000
Series - L2000
Subsystem - MC/ServiceGuard (SG)
RESOLUTION
You are running out of connections because you probably run a script that issues cmviewcl, or other ServiceGuard commands. Each command opens a connection to ServiceGuard's hacl-local port (5304/tcp). Since the number of cmcld's filedescriptors for connections of this kind is limited, there is a maximum number of ServiceGuard commands that can be executed at the same time. To avoid this problem, you can reduce the number of parallel ServiceGuard commands.
Another approach is to increase the ServiceGuard parameter MAX_CONFIGURED_PACKAGES. When you increase this parameter, you also increase the maximum number of local connections accepted by the cmcld. However, this leads to a more excessive use of system resources, and is therefore not the preferred solution. You can assign a maximum value of 30 to the MAX_CONFIGURED_PACKAGES parameter.
================================================
-- "The lyf so short, the craft so long to lerne." - Chaucer -------------------------------------------------- BV <mailto:BobVance@alumni.caltech.edu> Sr Technical Consultant Storage, Networking, HP-UX, Linux Inplexus - a Cnetics group 7076 Peachtree Ind. Blvd. Office 404-446-0404 Norcross, GA 30071 Fax 801-327-6152 Home Office 770-977-3476 ================================================== -----Original Message----- From: hpux-admin-owner@DutchWorks.nl [mailto:hpux-admin-owner@DutchWorks.nl]On Behalf Of Edi CAHYADI Sent: Wednesday, March 29, 2006 11:36 PM To: hpux-admin@dutchworks.nl Subject: [HPADM] cmcld: Stopped accepting local connection requests Hi Gurus, I found some messages from syslog: Mar 28 09:00:17 MHD003 cmcld: Stopped accepting local connection requests because there are currently too many concurrent connections to the daemon (20). Mar 28 09:00:17 MHD003 cmcld: There are no longer too many local connections. Now accepting local connection requests. Mar 28 09:00:18 MHD003 cmcld: Stopped accepting local connection requests because there are currently too many concurrent connections to the daemon (20). Mar 28 09:00:18 MHD003 cmcld: There are no longer too many local connections. Now accepting local connection requests. Mar 28 10:03:33 MHD003 cmcld: Stopped accepting local connection requests because there are currently too many concurrent connections to the daemon (20). Mar 28 10:03:33 MHD003 cmcld: There are no longer too many local connections. Now accepting local connection requests. I am wondering what had caused this error, and whether there really is any impact on the cluster. Regards, Edi -- ---> Please post QUESTIONS and SUMMARIES only!! <--- To subscribe/unsubscribe to this list, contact majordomo@dutchworks.nl Name: hpux-admin@dutchworks.nl Owner: owner-hpux-admin@dutchworks.nl Archives: ftp.dutchworks.nl:/pub/digests/hpux-admin (FTP, browse only) http://www.dutchworks.nl/htbin/hpsysadmin (Web, browse & search) -- No virus found in this incoming message. Checked by AVG Free Edition. Version: 7.1.385 / Virus Database: 268.3.3/296 - Release Date: 3/29/2006 -- No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.1.385 / Virus Database: 268.3.3/296 - Release Date: 3/29/2006 -- ---> Please post QUESTIONS and SUMMARIES only!! <--- To subscribe/unsubscribe to this list, contact majordomo@dutchworks.nl Name: hpux-admin@dutchworks.nl Owner: owner-hpux-admin@dutchworks.nl Archives: ftp.dutchworks.nl:/pub/digests/hpux-admin (FTP, browse only) http://www.dutchworks.nl/htbin/hpsysadmin (Web, browse & search)
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 11:02:52 EDT