>
>> If the hash is for general purposes, then it should not be encoded by base64
>> at the end. Array of bytes should be returned from the hash. If it's needed,
>> base64 can be done later on the returned array of bytes.
>>
>> BR,
>> Zoran
>>
>>
ponse(rde_amf_cb->amf_hdl, invocation, error);
TRACE_LEAVE2("rc = %u, error = %d", (unsigned) rc, (int) error);
}
regards,
Anders Widell
On 02/24/2017 12:34 PM, ramesh betham wrote:
> I don't think RED_UP always come first or later MDS_UP or vice-versa.
> Most l
n
Comments (indicate scope for each "y" above):
-
changeset f3551da7913d47d5f0a05291620febefb777c40d
Author: Anders Widell
Date: Fri, 24 Feb 2017 10:56:23 +0100
base: Add a collision-resistant hash function [#2266]
Add
README |2 +
src/base/Makefile.am| 17 --
src/base/hash.cc| 28 +++
src/base/hash.h | 67 +++
src/base/tests/hash_test.cc | 109
5 files changed, 217 ins
these remaining enhancement tickets so that they can be pushed before we
tag:
2267base: Add a support function for reading the FQDNAnders Widell
2266base: Add a hash functionAnders Widell
2259amf: support for si-swap admin op for N+M model. PraveenPraveen
2258log: add
n
Comments (indicate scope for each "y" above):
-
Re-based patch so that it can be applied on the latest changeset.
changeset fc5e51e2c1242a5fae9ddde8f4aa9d0b6532d8ec
Author: Anders Widell
Date: Tue, 28 Feb 2017 10:46:29 +0100
README |2 +
src/base/Makefile.am| 18 --
src/base/hash.cc| 28 +++
src/base/hash.h | 67 +++
src/base/tests/hash_test.cc | 109
5 files changed, 217 ins
are
pushed to the Mercurial repository before we open up the GIT repository.
Please prioritize the tickets on the 5.2.RC1 milestone, which as listed
at the following URL:
https://sourceforge.net/p/opensaf/tickets/milestone/5.2.RC1/
regards,
Anders Widell
n
Other y
Comments (indicate scope for each "y" above):
-
changeset a5632c11c700a9788d8c7d26b1de4b7c616ba050
Author: Anders Widell
Date: Mon, 06 Mar 2017 11:03:14 +0100
uml: Fix busybox build p
could in theory contain a valid handle - i.e.
some other handle that was initialized in a totally different part of
the code. So in short: if a function fails, then it is safest to not
make any assumption about the contents of output parameters from that
function.
regards,
Anders Widell
On 03/03
Ack with comment: You have to update opensaf.spec.in to insert
/sbin/shutdown into /etc/sudoers.
regards,
Anders Widell
On 03/06/2017 04:27 PM, Hans Nordeback wrote:
> scripts/opensaf_reboot | 2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
>
> diff
Ack.
regards,
Anders Widell
On 03/07/2017 08:20 AM, Hans Nordeback wrote:
> src/nid/nodeinit.cc | 2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
>
> diff --git a/src/nid/nodeinit.cc b/src/nid/nodeinit.cc
> --- a/src/nid/nodeinit.cc
> +++ b/src/nid/nod
One more comment: the syslog message on the line before the shutdown
command also has to be updated.
regards,
Anders Widell
On 03/07/2017 03:03 PM, Anders Widell wrote:
> Ack with comment: You have to update opensaf.spec.in to insert
> /sbin/shutdown into /etc/sudoers.
>
> regards
MDS_DIRECT_SEND, no matter if the MDS API call succeeds or
fails. Doesn't this patch change this behaviour in MDS?
regards,
Anders Widell
On 03/06/2017 09:00 AM, Hoang Vo wrote:
> src/base/sysf_mem.c| 3 +++
> src/mds/mds_c_sndrcv.c | 34 +++---
> 2
where it is used and modify the code appropriately. From what
I can see, the users of MDS_DIRECT_SEND currently don't free the
allocated memory if the MDS call fails.
regards,
Anders Widell
On 03/09/2017 03:39 AM, Vo Minh Hoang wrote:
>
> Dear Anders Widell,
>
> Yes, as you ment
src/base/conf.cc| 66 +---
src/base/conf.h | 5
src/nid/nodeinit.cc | 5
3 files changed, 62 insertions(+), 14 deletions(-)
Read the fully qualified domain name once and store it in the local file system.
This will solve prob
Other n
Comments (indicate scope for each "y" above):
-
changeset 3a44e8502e4343de1ecd5722a453e58f21c4364c
Author: Anders Widell
Date: Tue, 14 Mar 2017 12:31:17 +0100
base: Cache the FQDN to avoid problems caus
n be listed using the following link:
https://sourceforge.net/p/opensaf/tickets/search/?q=status%3A%28accepted+review+fixed%29+AND+_milestone%3A%285.2.FC+5.2.RC1+5.2.RC2+5.2.0%29+AND+_type%3Aenhancement&limit=100
thanks,
Ander
n
Comments (indicate scope for each "y" above):
-
changeset 93ed88041a574a307567e89b3ddc039e8c2a6a50
Author: Anders Widell
Date: Tue, 14 Mar 2017 16:13:30 +0100
build: Add missing README files to distribution tarb
python/Makefile.am | 4 +++-
src/amf/Makefile.am | 9 -
src/ckpt/Makefile.am | 7 ++-
src/clm/Makefile.am | 6 +-
src/dtm/Makefile.am | 3 +++
src/imm/Makefile.am | 4 +++-
src/log/Makefile.am | 8 +++-
src/ntf/Makefile.am | 8 +++-
src/plm/Makefile.am | 7
Ack.
regards,
Anders Widell
On 03/13/2017 01:16 PM, Hans Nordeback wrote:
> src/base/logtrace.c | 7 ++-
> 1 files changed, 6 insertions(+), 1 deletions(-)
>
>
> Even though not being LSB it would be good to use thread id instead of
> process id in trace.
>
&g
Other n
Comments (indicate scope for each "y" above):
-
changeset a0116a70ae42270f5aca07c29d2ecaf122ea7cd0
Author: Anders Widell
Date: Wed, 15 Mar 2017 14:59:43 +0100
build: Remove dependencies to internal OpenSAF
Makefile.am | 16 +-
README | 12 +--
samples/mqsv/Makefile.am | 4 ++
scripts/create_empty_library | 69
4 files changed, 90 insertions(+), 11 deletions(-)
Remove dependencies to internal Op
-blocking tickets can be pushed until the final
5.2.0 release. I have moved the date for the 5.2.RC2 tagging to Friday
2017-03-24 to give more time for these tickets.
regards,
Anders Widell
--
Check out the vibrant tech
tickets to the "next" milestone if possible (i.e. if
the problem doesn't have to be solved before the 5.2 release).
thanks,
Anders Widell
--
Check out the vibrant tech community on one of the world'
.hgignore |1 +
Makefile.am| 15 ++-
README |3 +-
src/base/Makefile.am |1 -
src/base/hash.cc | 179 +
src/base/hash.h| 37 ++--
sr
Comments (indicate scope for each "y" above):
-
changeset 5169d44e79a3852ca4ae7649d17904953fb0fc12
Author: Anders Widell
Date: Mon, 27 Mar 2017 08:33:16 +0200
base: Use dlopen to avoid LSB violation [#2396]
Ticket [#2266]
n
Comments (indicate scope for each "y" above):
-
changeset cd17b015441ce1b14db16bdb027a14634c2a0c6e
Author: Anders Widell
Date: Thu, 30 Mar 2017 14:18:58 +0200
base: Use getdomainname() to build the FQDN [#2405]
Queryi
src/base/conf.cc | 28 +++-
src/base/conf.h | 2 ++
2 files changed, 29 insertions(+), 1 deletions(-)
Querying the DNS can potentially be very slow, especially in the case when the
DNS servers are not responding e.g. due to a misconfiguration. Therefore, we
instead us
Thanks, I have fixed it now.
regards,
Anders Widell
On 03/31/2017 03:07 PM, Neelakanta Reddy wrote:
> Hi
>
> Failing in gcc 6.1(passed in gcc 4.9):
>
> CXXsrc/base/lib_libopensaf_core_la-hash.lo
> src/base/hash.cc:97:1: error:
> '{anonymous}::HashLibrary::~Hash
release. The full list of
5.2 enhancement tickets can be listed using the following link:
https://sourceforge.net/p/opensaf/tickets/search/?q=status%3A%28accepted+review+fixed%29+AND+_milestone%3A%285.2.FC+5.2.RC1+5.2.RC2+5.2.0%29+AND+_type%3Aenhancement&limit=100
thanks,
Anders Wi
n
Other n
Comments (indicate scope for each "y" above):
-
changeset 940295ebc4b1d4b67616fe6e80ee09b74c4ddcad
Author: Anders Widell
Date: Thu, 06 Apr 2017 13:14:01 +0200
ckpt:
src/ckpt/ckptnd/ckptnd.conf | 6 ++
src/ckpt/ckptnd/cpnd_cb.h | 3 ++-
src/ckpt/ckptnd/cpnd_init.c | 3 ++-
src/ckpt/ckptnd/cpnd_proc.c | 9 +
src/ckpt/ckptnd/cpnd_res.c | 9 +
5 files changed, 20 insertions(+), 10 deletions(-)
The setting OSAF_CKPT_SHM_ALLOC_GUARANTE
See reply below.
thanks,
Anders Widell
On 04/07/2017 05:17 AM, A V Mahesh wrote:
> Hi Anders Widell,
>
> Ack with following comments :
>
> On 4/6/2017 4:55 PM, Anders Widell wrote:
>> The setting OSAF_CKPT_SHM_ALLOC_GUARANTEE=0 results in increased
>> memory us
src/ckpt/ckptnd/cpnd_main.c | 18 ++
1 files changed, 18 insertions(+), 0 deletions(-)
The CKPT node director needs one file descriptor per open replica. Since the
maximum number of replicas (CPND_MAX_REPLICAS) is 1000, the CKPT node director
needs 1000 file descriptor just for
n
Comments (indicate scope for each "y" above):
-
changeset 2a19e5a90dce2082d7294ae8a56649fad65c221e
Author: Anders Widell
Date: Sat, 08 Apr 2017 10:17:12 +0200
ckpt: Increase limit for number of file desciptors
Mercurial and GIT during a short
transition period. A separate announcement will be sent out with more details
regarding how to use GIT and Mercurial in OpenSAF, as well as our new branching
strategy.
A big thanks to all who have contributed to this release!
regards,
Anders Widell
repository and review script, and report anything that is unclear or
isn't working.
regards,
Anders Widell
--
Check out the vibrant tech community on one of the world's
A question:
I noticed that clms_send_is_member_info() is called twice from
clms_imm_node_unlock(), but you only modified the first call (the one
without PLM). Shouldn't the PLM case also be updated? Is a similar
change also applicable in clms_imm_node_shutdown?
regards,
Anders Widell
Ack. Re-based version of the patch attached.
regards,
Anders Widell
On 04/03/2017 10:09 AM, praveen.malv...@oracle.com wrote:
src/clm/clmd/clms_evt.c | 2 +-
src/clm/clmd/clms_imm.c | 11 +--
src/clm/clmd/clms_mbcsv.c | 2 +-
src/clm/clmd/clms_util.c | 16
" ]; then
^-- SC2166: Prefer [ p ] || [ q ] as [
p -o q ] is not well defined.
Line 33:
if [ "$2" != "adm" -a "$2" != "mem" -a "$2" != "id" -a "$2" != "all" ]; then
Ack with comments:
1) Missing space between if statement and opening parenthesis.
2) Line line length exceeds 80 characters
3) The if statement should actually be removed - call SetPeerState()
unconditionally here.
regards,
Anders Widell
On 04/11/2017 03:10 PM, Zoran Milinkovic wrote
Hi!
Wouldn't this problem, as you already hint at, be solved if we simply
change the inotify condition to wait for IN_CLOSE_WRITE? That sound like
a better solution than introducing a retry loop here...
regards,
Anders Widell
On 04/20/2017 08:14 AM, Minh Chau wrote:
> At
Woundn't it be enough to wait for IN_CLOSE_WRITE only? Maybe we can
introduce a new method in FileNotify, instead of extending
WaitForFileCreation()?
/ Anders Widell
On 04/21/2017 03:49 PM, minh chau wrote:
> Hi Anders,
>
> It is because WaitForFileCreation() is also called in
Another approach could be to make sure that the pid file is created
atomically, e.g. by writing to a temporary file, closing it, and then
renaming it. Then we need to use inotify to wait for file rename operations.
regards,
Anders Widell
On 04/21/2017 04:12 PM, minh chau wrote:
> Hi And
->glnd_res_lock_mds_ctxt;$
src/lck/lcknd/glnd_restart.c:251: WARNING: please, no spaces at the
start of a line
+restart_res_lock_list_info->glnd_res_lock_mds_ctxt;$
src/lck/lcknd/glnd_restart.c:361: ERROR: code indent should use tabs
where possible
+node_id == m_NCS_NO
Ack.
regards,
Anders Widell
On 04/27/2017 03:55 PM, Hans Nordeback wrote:
> ---
> src/base/sysf_tmr.c | 20 +---
> 1 file changed, 17 insertions(+), 3 deletions(-)
>
> diff --git a/src/base/sysf_tmr.c b/src/base/sysf_tmr.c
> index d4bd8de12..42679e103 10064
te on the pid files (Ubuntu start-stop-daemon and LSB
start_daemon). Maybe the solution is as simple as using flock() when
reading the pid file?
regards,
Anders Widell
On 05/02/2017 03:10 PM, Hans Nordebäck wrote:
> Hi Minh,
>
> some minor comments below, I discussed with Anders a
as Hans says - otherwise
there can be a risk for denial of service.
regards,
Anders Widell
On 05/03/2017 08:52 AM, Hans Nordebäck wrote:
> Hi Minh,
>
> the flock is used in non blocking mode, not sure but as I mentioned
> earlier why not add a new function
>
> FileNotify::Wa
the file (maybe not needed - done by close?)
ENDIF
CLOSE the file
ENDIF
WAIT for inotify event
END REPEAT
regards,
Anders Widell
On 05/03/2017 10:46 AM, Anders Widell wrote:
> I think osaftransportd must do this sequence:
>
> * Open the file
>
>
Hi!
Will you update the documentation (README and PR doc) as well? Now that
we move towards continuous delivery we must be careful to keep the
documents up-to-date with the code at all times.
thanks,
Anders Widell
On 05/03/2017 06:37 AM, Gary Lee wrote:
> Hi Alex
>
> The value is
that by default enables this feature. But if
the object doesn't exist we should provide legacy behaviour.
regards,
Anders Widell
On 05/03/2017 04:24 PM, Alex Jones wrote:
> Hi Lennart,
>
>This was discussed at a recent TLC meeting, and it was agreed that any
> behavior that
, but at least the problem only exists when upgrading
from 5.2. It is easier to miss information from the release notes when
upgrading from an older release.
regards,
Anders Widell
On 05/04/2017 04:10 AM, Alex Jones wrote:
> One potential wrinkle I see with this
>
> The new AMF/SMF
where we create the pidfile atomically is the best one.
I will review the patch you sent out for review.
thanks,
Anders Widell
On 05/04/2017 01:54 PM, minh chau wrote:
> Hi Anders, Hans,
>
> I think there is still a deadlock even with nonblocking flock() in a
> below scenario
> -
Ack. Please squash the three commits into one single commit before pushing.
regards,
Anders Widell
On 05/09/2017 03:56 PM, Alex Jones wrote:
> saLckResourceOpen returns zero during controller failover
>
> Some functions in lckd are returning SaAisErrorT which is uninitialized.
>
Ack with comments, marked AndersW below.
regards,
Anders Widell
On 04/28/2017 11:36 AM, Minh Chau wrote:
> At startup, osaftransportd waits for osafdtmd.pid file creation
> and then reads dtm pid. If osafdtmd.pid has not been completedly
> created but osaftransportd still receives
Ack with one comment: --admin_op should be renamed to --admin-op (dash instead
of underscore as word separator).
regards,
Anders Widell
On 05/12/2017 11:54 AM, Praveen wrote:
> This is V2 which incorporates comments given on V1.
>
> clm-adm: for performing CLM admin operations on CLM
Add support for using DESTDIR to point to the installation root directory when
installing the OpenSAF sample programs.
---
samples/amf/proxy/Makefile.am| 6 +++---
samples/amf/sa_aware/Makefile.am | 8
samples/amf/wrapper/Makefile.am | 4 ++--
3 files changed, 9 insertions(+), 9 dele
revision 4b8b2ff831fb6ed818d400e30a7a035b70b77ad8
Author: Anders Widell
Date: Mon, 15 May 2017 12:31:07 +0200
osaf: Support DESTDIR when installing OpenSAF samples [#2460]
Add support for using DESTDIR to point to the installation root directory when
installing the OpenSAF sample programs.
Complet
before removing an existing pidfile. But for now,
I think it is enough to just remove it.
regards,
Anders Widell
On 05/16/2017 09:50 AM, minh chau wrote:
> Hi Ramesh,
>
> I'm going to revise the patch according to Anders and Hans' comments.
> Do you have any comments?
>
The exit codes for an init script are defined by LSB:
http://refspecs.linuxbase.org/LSB_3.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html
so if we are going to touch the init scipt, I think we ought to align
the exit codes to the above specification.
regards,
Anders Widell
On 05/17
ation repositories.
regards,
Anders Widell
--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_
Hi!
Can't you return 150 from the opensafd stop command instead? Then the
caller will know that he has to retry opensafd stop. Or maybe put the
retry inside the opensafd script?
/ Anders W
On 05/19/2017 12:46 PM, Rafael Odzakow wrote:
Internally opensafd creates a lock file during start/st
Ok I think 2. is a good solution. It is more generic and doesn't treat
opensafd in any special way compared to the other init scripts that are
installed on the system.
/ Anders Widell
On 06/01/2017 10:52 AM, Rafael Odzakow wrote:
1. Not possible, check my other response.
2. I am OK
$timeout - $interval`
^-- SC2006: Use $(..) instead of legacy `..`.
^-- SC2003: expr is antiquated. Consider rewriting
this using $((..)), ${} or [[ ]].
thanks,
Anders Widell
On 06/01/2017 02:37 PM, Rafael Odzakow wrote:
Internally opensafd creates a lo
if we could add a SA_CLM_AF_TIPC = 3, but if you don't
have the time to do that now it is probably better to keep the
attributes empty (as before) when using TIPC.
regards,
Anders Widell
On 05/26/2017 08:54 AM, Praveen wrote:
CLM gets ip address and address family from MDS in node up
Ack. Not sure if all this retry logic is needed - especially not for the
EBUSY case - but let's keep it for now.
regards,
Anders Widell
On 05/25/2017 06:52 AM, Minh Chau wrote:
At startup, osaftransportd waits for osafdtmd.pid file creation
and then reads dtm pid. If osafdtmd.pid ha
Note that we maybe should avoid adding new a dependency towards the
inotifywait tool...
regards,
Anders Widell
On 06/02/2017 02:38 PM, Rafael Odzakow wrote:
Thanks I am trying to implement the suggestions and will send out a
new patch soon.
On 06/02/2017 08:51 AM, Hans Nordebäck wrote
copy into memory allocated
by calloc()). The function proc_node_up_msg() contains another memcpy
where the string is copied form IPADDR to CLMS_CLUSTER_NODE, and this
piece of code doesn't guarantee NUL-termination either. Please check
that code guarantees NUL-termination.
regards,
Anders
servicesy
Core libraries n
Samples n
Tests n
Other n
Comments (indicate scope for each "y" above):
-
revision 312674e5ca8bb4ec50a1d2d843d3db3bcbddfe02
Author: Anders Widell
D
Ensure compatibility with Python 3 by running the Python source code files
through the Automated Python 2 to 3 code translation tool "2to3". For more
information about this tool, see:
https://docs.python.org/3/library/2to3.html
---
python/pyosaf/saEnumConst.py | 6 +++---
python/pyos
Ack.
regards,
Anders Widell
On 06/07/2017 09:13 AM, Hoang Vo wrote:
When a test case call test_validate() multiple times, the test report
is not clear so enhance it for easier recognizing result.
---
src/osaf/apitest/utest.c | 33 -
1 file changed, 24
Hi!
If there are no comments, I will push this by Friday.
regards,
Anders Widell
On 05/15/2017 12:48 PM, Anders Widell wrote:
Summary: osaf: Support DESTDIR when installing OpenSAF samples [#2460]
Review request for Ticket(s): 2460
Peer Reviewer(s): Ramesh
Pull request to:
Affected branch
revision e53a35b9875841a5336fa6eef312332fa786568c
Author: Anders Widell
Date: Mon, 19 Jun 2017 15:57:54 +0200
osaf: Fix build problems when using GCC 7.1.1 [#2502]
Fix build problems caused by the following warnings:
* -Werror=format-overflow=
* -Werror=parentheses
* -Werror=unused-variable
Complete diffstat:
-
Fix build problems caused by the following warnings:
* -Werror=format-overflow=
* -Werror=parentheses
* -Werror=unused-variable
---
src/amf/amfd/imm.cc| 2 -
src/clm/apitest/tet_ClmLongRdn.c | 24 +---
src/clm/apitest/tet_ClmOiOps.c
e or directory
opensafd start/stop already in progress. Unable to continue ... (warning).
To forcefully start/stop OpenSAF remove
/var/lock/subsys/opensafd_inprogress ... (warning).
I think you need to make sure that /var/lock/subsys is created before
mutex_create() is called.
regards,
Anders
revision 6299580109043b8831d53cd60d5cc10cb66901d4
Author: Anders Widell
Date: Thu, 22 Jun 2017 15:46:11 +0200
build: Move all AIS C header files to a common subdirectory under src [#2472]
This simplifies the directory structure and build system by collecting all AIS
header files into a separate subtre
Hi!
Did you get a chance to look at tickets [#2502] and [#2472] yet?
If there are no comments I would like to push them by the end of this week.
thanks,
Anders Widell
On 06/20/2017 10:57 AM, Anders Widell wrote:
Summary: osaf: Fix build problems when using GCC 7.1.1 [#2502]
Review request
Ack.
regards,
Anders Widell
On 06/28/2017 01:39 PM, Zoran Milinkovic wrote:
When active or standby nodes are introduced with request message, there is no
need to wait more for requesting the active role.
When standby node is introduced, then we are sure that there is an active node
agent libraries. Please use saAisNameBorrow() and
strlen(saAisNameBorrow()) instead.
* Inconsistent use of EXIT_FAILURE/EXIT_SUCCESS and 1/0 for exit() and
return from main (sometimes name is used, sometimes number).
regards,
Anders Widell
On 07/14/2017 11:02 AM, Praveen wrote:
Add a u
logging to syslog.
regards,
Anders Widell
On 07/20/2017 01:36 AM, jonas arndt wrote:
And now with a subject :-)
Hi Guys,
This is with regards to debianizing OpenSAF, i.e. ticket 2427.
As I go through the RPM dependencies in the opensaf.spec file I am
starting to think that there are some issues
other.
regards,
Anders Widell
On 07/20/2017 11:52 AM, jonas arndt wrote:
And if IMM is not there? What does AMF do then😀
I can have separate packages for separate services if that is
necessary. The issue is that the current RPM spec file is not
reflecting the correct dependencies. If the AMF libra
Hi!
Yes I have taken over as maintainer for BASE.
dpkg-shlibdeps seems like a good tool that finds valid problems. See
comments below, marked AndersW>
regards,
Anders Widell
On 07/23/2017 04:23 AM, jonas arndt wrote:
Hi Anders
As Ramesh left are you now the maintainer for Core (B
Ticket [#1453] added support for autoscaling, which allows scale-out from an
initial cluster containing at least one node. This commit adds support for
scaling out from a cluster containing zero nodes, or alternatively, a cluster
where the active node is not a configured node. The use cases are as
Avoid spamming the syslog with more than one log message in case CLM returns
SA_AIS_ERR_UNAVAILABLE (i.e. we are running on a currently unconfigured node).
---
src/amf/amfd/clm.cc | 6 --
src/amf/amfnd/clm.cc | 6 --
2 files changed, 8 insertions(+), 4 deletions(-)
diff --git a/src/amf/a
2ce
Author: Anders Widell
Date: Mon, 24 Jul 2017 15:41:51 +0200
clm: Make it possible for a node to scale out itself using autoscaling [#2509]
Ticket [#1453] added support for autoscaling, which allows scale-out from an
initial cluster containing at least one node. This commit adds support for
scalin
Re-try initializing the CLM API when it returns SA_AIS_ERR_UNAVAILABLE, so that
the NTF service properly waits for the node to become configured by the
autoscaling functionality.
---
src/ntf/ntfd/ntfs_clm.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/ntf/ntfd/ntfs_clm
Re-try initializing the CLM API when it returns SA_AIS_ERR_UNAVAILABLE. This
error code is returned if the LOG service has been started on an unconfigured
node, which may happen for a while when the autoscaling feature is used.
---
src/log/logd/lgs_clm.cc | 3 ++-
1 file changed, 2 insertions(+),
OpenSAF servicesn
Core libraries y
Samples n
Tests n
Other n
Comments (indicate scope for each "y" above):
-
revision 10a76da49f211674579fe5e03fe9d72c44e63709
Auth
The osaf_get_boot_time function returns the time stamp when the node was booted.
The osaf_timespec_average function returns the average of two time stamps.
---
src/base/Makefile.am | 12 ++-
src/base/osaf_time.c | 24 +
src/base/osaf_time.h
The node join request message now has two new fields: boot time and node
address. This allows us to provide more accurate and correct information in the
CLM node runtime attributes in the information model:
* The boot time field transmits the node's actual boot time to the CLM server.
Previously
ill out on review.
revision 56e7a62729974e03f3cd441a182121ea67937136
Author: Anders Widell
Date: Mon, 31 Jul 2017 14:02:07 +0200
clm: Include boot time and node address in join request message [#2489]
The node join request message now has two new fields: boot time and node
address. This allows us
Provide the node address as a command-line parameter when calling the scale-out
script. This can be useful if the scale-out script needs to contact the node
(e.g. copy some files to it or update some configuration on the node's local
disk) as part of the scale-out operation.
---
src/clm/clmd/clms_
n
OpenSAF servicesy
Core libraries n
Samples n
Tests n
Other n
Comments (indicate scope for each "y" above):
-
revision 054ddd10b50baaee87f686c7c7813f6d65ddd8d5
Auth
uot; above):
-
revision 0e40cc0bdab8a86cd04a0fea0a8b46409368375f
Author: Anders Widell
Date: Wed, 2 Aug 2017 12:40:44 +0200
rde: Avoid sending messages to peers that are not up [#2488]
Replace the request-reply message exchange between RDE peers with a simpler
protocol that just sends
Replace the request-reply message exchange between RDE peers with a simpler
protocol that just sends the reply message just after receiving peer up
notifications. This will minimize the probability of trying to send to a peer
that is down. Also remove the retransmission logic that was mostly needed
Ack.
regards,
Anders Widell
On 08/02/2017 10:13 AM, Rafael Odzakow wrote:
Using echo -n in opensafd causes delay write to log in a systemd
environment causing unconsistent order of the logs. "Starting opensaf"
will end up after "Startup finished" in the system log.
---
Hi!
Have you pushed this yet? I tried it and I think it works fine now!
regards,
Anders Widell
On 07/28/2017 01:44 PM, praveen malviya wrote:
Hi Anders,
Attached is the patch after incorporating the comments.
I will be pushing it on Monday.
Please go through it.
Thanks
Praveen
On 18-Jul
Divide the safe reboot in the opensaf_reboot script into two phases: in the
first phase we stop important OpenSAF services (especially IMM), and in the
second phase we reboot the nodes.
---
scripts/opensaf_reboot | 16 ++--
1 file changed, 14 insertions(+), 2 deletions(-)
diff --git a
uot; above):
-
revision f6f53573ee25913d99629e7e41cddf520085a776
Author: Anders Widell
Date: Wed, 2 Aug 2017 13:07:19 +0200
osaf: Divide the safe reboot into two phases [#2542]
Divide the safe reboot in the opensaf_reboot script into two phases: in the
first phase we stop important OpenSA
401 - 500 of 993 matches
Mail list logo
