Steve mentioned that strlen is appearing a little too high
in profiling, so this removes one redundant use from what is
probably a hot code path.
I will do the same to the else block next.
From 51eaaa809ac94ec20ddcbe62c651735106128a31 Mon Sep 17 00:00:00 2001
From: Jim Meyering
Dear Steve:
I had used something like your patch to test for awhile and it worked.
So, I think this patch will work and I will try it later.
During the investigation, I found something strange.
For example,
1. throw away active, assembly-last_frag_num == Y
2. mcast-fragmented == 1 or 0,
Here is an example CPG application which shows how to replicate state
between multiple processes on multiple machines. It is just a start,
but have a look. If you have interest in CPG, it may help you
understand the basics of the API.
Interesting, especially the use of 2 process groups.
Hi,
several times it's been reported that make install will override
the installed configuration.
This is wrong.
Instead of shipping/installing corosync.conf (that in 99.9% of the case
it needs to be customized), ship it as corosync.conf.example and let the
user merge what they need.
The patch
This patch fixes one exit path error handling in logsys_format_set
and add 2 error checks when returning from
logsys_config_file_set_unlocked.
Fabio
Index: exec/logsys.c
===
--- exec/logsys.c (revision 2093)
+++ exec/logsys.c
On Mon, 2009-04-20 at 11:18 +0200, Jim Meyering wrote:
Fabio M. Di Nitto wrote:
several times it's been reported that make install will override
the installed configuration.
This is wrong.
Instead of shipping/installing corosync.conf (that in 99.9% of the case
it needs to be
Fabio M. Di Nitto wrote:
Hi,
several times it's been reported that make install will override
the installed configuration.
This is wrong.
Instead of shipping/installing corosync.conf (that in 99.9% of the case
it needs to be customized), ship it as corosync.conf.example and let the
Fabio M. Di Nitto wrote:
This patch fixes one exit path error handling in logsys_format_set
and add 2 error checks when returning from
logsys_config_file_set_unlocked.
Fabio
Index: exec/logsys.c
===
...
@@ -1248,13 +1253,14
On Mon, 2009-04-20 at 11:13 +0100, Chrissie Caulfield wrote:
Fabio M. Di Nitto wrote:
Hi,
several times it's been reported that make install will override
the installed configuration.
This is wrong.
Instead of shipping/installing corosync.conf (that in 99.9% of the case
it
Here are two warnings:
schedwrk.c:52: warning: cast from pointer to integer of different size
schedwrk.c:110: warning: cast to pointer from integer of different size
Here's how to avoid them.
These days, the only safe way is via a union.
The two accessor functions are just so that the code
Hello!
I have some questions about using openais checkpoints
I have two nodes with openais. One node is active and one is passive. I
have enabled amf and ckpt service.
The node active write data in checkpoints. When it crashes, the passive
becomes active and reads the good data.
It works
Starting from this warning (1 of 4 identical):
confdb.c:559: warning: passing argument 4 of 'api-object_find_create'
from incompatible pointer type
confdb.c:559: note: expected 'hdb_handle_t *' but argument is of type
'mar_uint64_t *'
Here's the offending code in
looks good for merge
On Mon, 2009-04-20 at 10:23 +0200, Fabio M. Di Nitto wrote:
Hi,
several times it's been reported that make install will override
the installed configuration.
This is wrong.
Instead of shipping/installing corosync.conf (that in 99.9% of the case
it needs to be
That bug is fixed in current whitetank tip. Which version are you
using?
regards
-steve
On Mon, 2009-04-20 at 13:50 +0200, Ludovic Lemarchand wrote:
Hello!
I have some questions about using openais checkpoints
I have two nodes with openais. One node is active and one is passive. I
have
good for merge.
Regards
-steve
On Mon, 2009-04-20 at 12:29 +0200, Jim Meyering wrote:
Here are two warnings:
schedwrk.c:52: warning: cast from pointer to integer of different size
schedwrk.c:110: warning: cast to pointer from integer of different size
Here's how to avoid them.
These
This change removes some useless if-before-free tests.
FYI, it was all done mechanically.
For rationale, portability discussion, etc, see these:
http://thread.gmane.org/gmane.comp.version-control.git/74187
http://thread.gmane.org/gmane.comp.lib.gnulib.bugs/12712
Steven Dake writes:
Some people on these lists may be interested to know what is happening
with Corosync and OpenAIS ABIs as well as our schedules for 1.0.
what is corosync and how is it related to openais? the figure on page
http://www.clusterlabs.org/wiki/Main_Page
does not contain a
On Fri, Apr 17, 2009 at 10:56:47PM -0700, Steven Dake wrote:
On Sat, 2009-04-18 at 07:49 +0200, Dietmar Maurer wrote:
like a 'merge' function? Seems the algorithm for checkpoint recovery
always uses the state from the node with the lowest processor id?
Yes that is right.
So if
On Sat, Apr 18, 2009 at 07:49:12AM +0200, Dietmar Maurer wrote:
like a 'merge' function? Seems the algorithm for checkpoint recovery
always uses the state from the node with the lowest processor id?
Yes that is right.
So if I have the following cluster:
Part1: node2 node3 node4
Steven Dake wrote:
good for merge.
On Mon, 2009-04-20 at 12:29 +0200, Jim Meyering wrote:
Here are two warnings:
schedwrk.c:52: warning: cast from pointer to integer of different size
schedwrk.c:110: warning: cast to pointer from integer of different size
Thanks.
committed
On Sat, Apr 18, 2009 at 09:37:26AM +0200, Dietmar Maurer wrote:
Yes, forcing the losers to reset and start from scratch is a must, but we
end up doing that a layer above corosync. That means the losers often
reappear again through corosync/cpg prior to being forced out.
Are you talking
This fixes the last of the unchecked-strdup problems:
From 143bf3681c180c0dd8c14596fde240808827615a Mon Sep 17 00:00:00 2001
From: Jim Meyering meyer...@redhat.com
Date: Mon, 20 Apr 2009 18:51:46 +0200
Subject: [PATCH] corosync-cfgtool.c: handle strdup failure gracefully
*
On Sat, Apr 18, 2009 at 03:55:57AM -0700, Steven Dake wrote:
On Sat, 2009-04-18 at 12:47 +0200, Dietmar Maurer wrote:
At least the SA Forum does not mention such strange behavior. Isn't that
a serious bug?
Yes, I'd consider it a serious bug.
Consider 2 Partitions with one checkpoint:
see subject.
Index: configure.ac
===
--- configure.ac (revision 1831)
+++ configure.ac (working copy)
@@ -101,7 +101,8 @@
AC_CHECK_FUNCS([alarm alphasort atexit bzero dup2 endgrent endpwent fcntl \
getcwd getpeerucred gettimeofday
Those checks are there for a reason. Did you test your patch? Often
those data sets are NULL.
Regards
-steve
On Mon, 2009-04-20 at 16:49 +0200, Jim Meyering wrote:
Same thing, but for openais:
From 41a09aa652374e5dc89729517d031303f90b1990 Mon Sep 17 00:00:00 2001
From: Jim Meyering
no these are not useless.
Regards
-steve
On Mon, 2009-04-20 at 16:48 +0200, Jim Meyering wrote:
This change removes some useless if-before-free tests.
FYI, it was all done mechanically.
For rationale, portability discussion, etc, see these:
looks good
On Tue, 2009-04-21 at 05:45 +0200, Fabio M. Di Nitto wrote:
Hi Steven,
the patch in attachment fixes the problem reported by Ryan of logfiles
not having newlines and skip an unnecessary if ...
the problem was a misplaced set of () in the output filter check on all
3
Hi guys,
in order to match the new logging config spec, 2 logging config keywords
had to be changed.
The following patch adds the old keywords as compatibility layer and
also add a warning that the config keyword has been obsoleted.
If both old and new keyword are present, the new one is
Hi,
Andrew requested an API interface to be able to flush logs on demand.
Add logsys_flush call to signal the logging thread to do its job.
Fabio
--- include/corosync/engine/logsys.h (revision 2100)
+++ include/corosync/engine/logsys.h (working copy)
@@ -150,6 +150,8 @@
/*
* External API -
Hi Steven,
this patch makes sure that logs are flushed when we are about to die.
In most cases logsys is not fully initialized yet and all the error logs
are just queued in the buffers. We want to guarantee that the user can
see them.
Fabio
--- exec/util.c (revision 2100)
+++ exec/util.c
looks good for merge
On Tue, 2009-04-21 at 06:15 +0200, Fabio M. Di Nitto wrote:
Hi,
Andrew requested an API interface to be able to flush logs on demand.
Add logsys_flush call to signal the logging thread to do its job.
Fabio
___
Openais
looks good
does logsys_fork_completed() have a check to make sure it hasn't already
been run?
Regards
-steve
On Tue, 2009-04-21 at 06:24 +0200, Fabio M. Di Nitto wrote:
Hi Steven,
this patch makes sure that logs are flushed when we are about to die.
In most cases logsys is not fully
Committed revision 2102.
Fabio
On Mon, 2009-04-20 at 21:27 -0700, Steven Dake wrote:
looks good
On Tue, 2009-04-21 at 06:06 +0200, Fabio M. Di Nitto wrote:
Hi guys,
in order to match the new logging config spec, 2 logging config keywords
had to be changed.
The following patch
Committed revision 2104.
Fabio
On Mon, 2009-04-20 at 21:28 -0700, Steven Dake wrote:
looks good
does logsys_fork_completed() have a check to make sure it hasn't already
been run?
Regards
-steve
On Tue, 2009-04-21 at 06:24 +0200, Fabio M. Di Nitto wrote:
Hi Steven,
this patch
Merging/reconciling a partition is *fundamentally* not possible for
any
application that requires virtual synchrony. You cannot rewrite
history,
But I can re-sync the state.
and VS is all about event history. It has nothing to do with
mechanism or
design, it has everything to do with the
The key point is that two cpg members (A and B) with different event
histories
will have inconsistent state, assuming that the state is derived from
the
history. So, after they are partitioned and merged, A needs to
replace
its
own state/history with B's or v.v.
Right. That is the idea of
36 matches
Mail list logo