[Pacemaker] error: qb_ipcs_us_connection_acceptor: Could not accept client connection: Too many open files (24)

[David Vossel]

----- Original Message -----
> From: "Nikola Ciprich" <nikola.ciprich at linuxbox.cz>
> To: pacemaker at oss.clusterlabs.org
> Sent: Tuesday, August 6, 2013 5:13:02 AM
> Subject: [Pacemaker] error: qb_ipcs_us_connection_acceptor: Could not accept client connection: Too many open files
> (24)
> 
> Hi,
> 
> I'd like to ask whether somebody met similar bug:
> 
> On one of the test two node clusters, node suddenly hung, and cib started
> spawning following messages:
> 
> error: qb_ipcs_us_connection_acceptor: Could not accept client connection:
> Too many open files (24)
> 
> in lsof, I see over thousand of opened /dev/shm files:

What version of libqb do you have installed? If you can upgrade libqb

-- Vossel

> 
> cib        5737 hacluster  DEL       REG               0,14
> 2615869 /dev/shm/qb-cib_rw-control-5737-25733-179
> cib        5737 hacluster  DEL       REG               0,14
> 2545021 /dev/shm/qb-cib_rw-control-5737-4605-178
> cib        5737 hacluster  DEL       REG               0,14
> 2410274 /dev/shm/qb-cib_rw-control-5737-1925-180
> cib        5737 hacluster  DEL       REG               0,14
> 2545640 /dev/shm/qb-cib_rw-control-5737-8828-177
> cib        5737 hacluster  DEL       REG               0,14
> 2495467 /dev/shm/qb-cib_rw-control-5737-2054-174
> cib        5737 hacluster  DEL       REG               0,14
> 2434602 /dev/shm/qb-cib_rw-control-5737-8659-176
> 
> 
> and also sockets:
> 
> cib        5737 hacluster 1003u     unix 0xffff880eaefee000      0t0
> 13885836 socket
> cib        5737 hacluster 1004u     unix 0xffff880eada76000      0t0
> 13849634 socket
> cib        5737 hacluster 1005u     unix 0xffff880eb37e7400      0t0
> 13847814 socket
> cib        5737 hacluster 1006u     unix 0xffff88099c120400      0t0
> 13866356 socket
> cib        5737 hacluster 1007u     unix 0xffff880eb7764000      0t0
> 13911546 socket
> cib        5737 hacluster 1008u     unix 0xffff880a7f579400      0t0
> 13847938 socket
> cib        5737 hacluster 1009u     unix 0xffff880a7f57e000      0t0
> 13883333 socket
> 
> OS is latest centos6 (RHEL6 clone), running x86_64 3.0.87 kernel
> 
> another important packages:
> 
> pacemaker-1.1.8-7.el6.x86_64
> cluster-glue-1.0.5-6.el6.x86_64
> clusterlib-3.0.12.1-49.el6.x86_64
> 
> Any idea on what this could be? Is this some known bug?
> 
> with best regards
> 
> nik
> 
> 
> 
> --
> -------------------------------------
> Ing. Nikola CIPRICH
> LinuxBox.cz, s.r.o.
> 28.rijna 168, 709 00 Ostrava
> 
> tel.:   +420 591 166 214
> fax:    +420 596 621 273
> mobil:  +420 777 093 799
> www.linuxbox.cz
> 
> mobil servis: +420 737 238 656
> email servis: servis at linuxbox.cz
> -------------------------------------
> 
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> 
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
>