[Openais] Cluster failure revisited, after a network attack

2011-08-23 Thread Jed Smith
Good morning, I am running Corosync 1.4.0 with Pacemaker 1.0.11. I've had a reappearance of the issue that I brought to the list before. This time, however, it was triggered by external stimulus. We use Corosync in a fairly high-pressure environment. Currently in the cluster in question I have 5

Re: [Openais] Cluster failure revisited, after a network attack

2011-08-23 Thread Jed Smith
Again, apologies for the additional message, thought I didn't have core files. Backtraces of cores attached. -- Jed Smith j...@jedsmith.org Thread 12 (Thread 10630): #0 0xf57fe416 in __kernel_vsyscall () #1 0xb76f8245 in sem_wait@@GLIBC_2.1 () from /lib/tls/i686/cmov/libpthread.so.0 #2 0xb7711

[Openais] Corosync 2.0 Feature Request: Support more RRP modes

2011-08-23 Thread Jiaju Zhang
As Corosync RRP only supports two modes (Active and Passive) which were described in that Totem redundant ring paper for now, can we support more modes like active-standby or proportional distributing the workload according to the bandwidth? For the active-standby mode, I think some users may have

[Openais] Corosync 2.0 Feature Request: Implement Totem Multiple-Ring Protocol

2011-08-23 Thread Jiaju Zhang
I remembered there is this feature in previous roadmap (might be in 3.0) from corosync.org. But I can't find this feature in the latest roadmap. Do we also have the plan to implement the totem multiple-ring in the future? Thanks, Jiaju ___ Openais mailin

Re: [Openais] aisexec crashes with SIGABRT

2011-08-23 Thread Christopher A. Kirke
Steve, appreciate the quick response, only happened to be running strace during one of the crashes. i've updated /etc/init.d/openais to enable core dump and changed openais.conf to run as root:asterisk instead of asterisk:asterisk so /var/lib/openais is available for writing. will post gdb output

Re: [Openais] Cluster failure revisited, after a network attack

2011-08-23 Thread Digimer
On 08/23/2011 09:12 AM, Jed Smith wrote: > Good morning, > > I am running Corosync 1.4.0 with Pacemaker 1.0.11. I've had a > reappearance of the issue that I brought to the list before. This > time, however, it was triggered by external stimulus. > > We use Corosync in a fairly high-pressure enviro

[Openais] [PATCH] Remove hdb.h header includes from unnecessary files

2011-08-23 Thread Steven Dake
The files in this patch do not use the hdb.h header. Signed-off-by: Steven Dake --- exec/totemrrp.c |1 - exec/totemsrp.c |1 - exec/totemudp.c |1 - exec/totemudp.h |1 - exec/totemudpu.c |1 - exec/totemudpu.h |1 - 6 files changed, 0 insertions(+), 6 deletions(-)

[Openais] [PATCH] Get rid of hdb usage in totempg.h interface

2011-08-23 Thread Steven Dake
hdb has some expense and is not necessary in the totempg.so runtime. This patch removes the dependence on hdb and instead uses a direct pointer. Signed-off-by: Steven Dake --- exec/main.c |2 +- exec/sync.c |2 +- exec/syncv2.c

Re: [Openais] [PATCH] Remove hdb.h header includes from unnecessary files

2011-08-23 Thread Angus Salkeld
On Tue, Aug 23, 2011 at 01:03:48PM -0700, Steven Dake wrote: > The files in this patch do not use the hdb.h header. ACK > > Signed-off-by: Steven Dake > --- > exec/totemrrp.c |1 - > exec/totemsrp.c |1 - > exec/totemudp.c |1 - > exec/totemudp.h |1 - > exec/totemudpu.c |

Re: [Openais] [PATCH] Get rid of hdb usage in totempg.h interface

2011-08-23 Thread Angus Salkeld
On Tue, Aug 23, 2011 at 02:13:07PM -0700, Steven Dake wrote: > hdb has some expense and is not necessary in the totempg.so runtime. This > patch removes the dependence on hdb and instead uses a direct pointer. ACK > > Signed-off-by: Steven Dake > --- > exec/main.c |2