Do you think to use "type:provider" as the key of the hash table is an
acceptable solution?
Andrew Beekhof wrote:
On 10/20/06, Huang Zhen <[EMAIL PROTECTED]> wrote:
Hi,
get_all_type_metadata() is to return metadata of all types of all RAs.
^^^
Till now, there is n
nux-ha.org/
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http
Andrew Beekhof wrote:
On 10/19/06, Huang Zhen <[EMAIL PROTECTED]> wrote:
The TSA plugin stopped the "make dist" because it did not give the
full path of source file in Makefile.am.
So "make dist" tried to find those files in the build directory.
It has been fixed in
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing
Ok, done.
Andrew Beekhof wrote:
On 9/28/06, Huang Zhen <[EMAIL PROTECTED]> wrote:
Hi,
The code related to the 3rd part quorum server has been pushed to the
hg.linux-ha.org.
And a document about quorum server is at
http://wiki.linux-ha.org/QuorumServerGuide now.
Any comment or suggest
Lars Marowsky-Bree wrote:
On 2006-10-02T21:03:14, Huang Zhen <[EMAIL PROTECTED]> wrote:
2_0_8 is the protocol version. Different versons of heartbeat may
share one same protocol. When we add a new implementation, we just
need put a new plugin with the new version. And when the
turn out the enlightenment?" - TISM
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Te
Andrew Beekhof wrote:
On 10/1/06, Huang Zhen <[EMAIL PROTECTED]> wrote:
Andrew Beekhof wrote:
> zhenh, can you please explain this change:
> http://hg.linux-ha.org/dev?cs=eb28a7552991
>
> there is nothing in that file that needs to be generated or filled
in by
> c
s. Is there a specific reason why
we're wasting one of hg's most powerful features?
It looks that I am not familiar with these powereful features of hg very well.
I may read the guide of hg again.
Sincerely,
Lars Marowsky-Brée
--
Best Regards,
Huang Zhen
Linux-HA, Linux Tec
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Be
reaker.
Besides, in general I'm thinking that doing the external quorum server
is a misjudgement of priorities, but, as it wasn't my call nor yours, I
won't elaborate on that ;-) (And, with the features in the PE, I've my
own featuritis to answer for, so I'm afraid I c
Hi,
The code related to the 3rd part quorum server has been pushed to the
hg.linux-ha.org.
And a document about quorum server is at
http://wiki.linux-ha.org/QuorumServerGuide now.
Any comment or suggestion on both the code or document are welcome.
Thanks!
--
Best Regards,
Huang Zhen
Linux-HA
nux-ha.org/
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
___
Linux-HA-Dev: Linux-HA-Dev@lists.lin
.
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
E
Patches are certainly welcome. ;-)
Sincerely,
Lars Marowsky-Brée
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
___
/*do we have quorum?*/
Sincerely,
Lars Marowsky-Brée
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
___
hanks for you comments.
Lars Marowsky-Bree wrote:
On 2006-08-11T16:48:55, Huang Zhen <[EMAIL PROTECTED]> wrote:
if (rc != EXECRA_OK || debug_level > 0) {
^^
Actually, all these logs are protected to only occur while debugging
alre
return cl_strdup(MSG_FAIL"\nwrong parameter
number"); \
___
Linux-ha-cvs mailing list
Linux-ha-cvs@lists.linux-ha.org
http://lists.community.tummy.com/mailman/listinfo/linux-ha-cvs
___
L
doesn't change anything.
It's "||" instead of "&&" in the condition.
So as the same reason above, it would log everytime that monitoring a stopped
resource.
In particular in error legs, we want logging.
Sincerely,
Lars Marowsky-Brée
--
Best Regards
ime':
151.1011700630188, 'min_time': 73.956845998764038}
Aug 11 13:08:00 Test ResourceRecover: {'elapsed_time': 2006.3324725627899,
'skipped': 0, 'calls': 49, 'success': 49, 'auditfail': 0, 'failure': 0,
'max_time
cess': 40, 'auditfail': 0, 'failure': 0,
'max_time': 89.774415016174316, 'min_time': 15.526778936386108}
Aug 09 05:55:14 Test SpecialTest1: {'elapsed_time': 4948.8167071342468,
'skipped': 0, 'calls': 49, 'success
TH_PROGS(POD2MAN, pod2man)
"
?
AC_PATH_PROGS(ROUTE, route)
AC_PATH_PROGS(RPM, rpmbuild)
if test x"${RPM}" = x""; then
___
Linux-ha-cvs mailing list
Linux-ha-cvs@lists.linux-ha.org
http://lists.community.tummy.com/mailman/listinfo/li
p://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
__
Andrew Beekhof wrote:
On 6/28/06, Huang Zhen <[EMAIL PROTECTED]> wrote:
Andrew,
You made several mistakes here, I would like to point out them.
MISTAKE 1 : Empty group is invalid configuration
Current crm-1.0.dtd shows :
So it's your code wrong instead of GUI.
are you trying
Lars Marowsky-Bree wrote:
On 2006-06-28T09:18:36, Huang Zhen <[EMAIL PROTECTED]> wrote:
You made several mistakes here, I would like to point out them.
MISTAKE 1 : Empty group is invalid configuration
Current crm-1.0.dtd shows :
So it's your code wrong instead of GUI.
Uhm. No
Lars Marowsky-Bree wrote:
On 2006-06-27T15:46:25, Huang Zhen <[EMAIL PROTECTED]> wrote:
Here is the design issues of the quorum server. Please review it.
http://wiki.linux-ha.org/QuorumServerDesign
1) I'd not call that "subcluster" when you mean "site"
e.
The various hacks that are currently in place have a very limited lifespan and
will be reverted.
--- You are receiving this mail because: ---
You are the assignee for the bug, or are watching the assignee.
You are on the CC list for the bug, or are watching someone who is.
--
Best
Hi,
Here is the design issues of the quorum server. Please review it.
http://wiki.linux-ha.org/QuorumServerDesign
Thanks!
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [E
Sorry for the mistake :)
I mean support more than two nodes.
One is in primary and others in secondary.
Lars Marowsky-Bree wrote:
On 2006-06-26T10:16:38, Huang Zhen <[EMAIL PROTECTED]> wrote:
The heartbeat can support more than one nodes, if drbd can support this
will be very h
Hi,
The heartbeat can support more than one nodes, if drbd can support this will be
very helpful.
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTE
e CIB and restore it. Maybe we need a CIB backup/restore feature
or tool.
(Be it for building another bunch of clusters from a script or for
reproducing a bug.)
Sincerely,
Lars Marowsky-Brée
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technol
t gets better.
This is a complex issue. I think it needs more thought, some
contemplation, some inspiration, and a picture of how the whole should
work.
We aren't going to get it here. But, of course, you should feel free to
keep on flogging this dead horse anyway ;-).
--
Best Reg
keep flogging this horse until it gets better.
Sincerely,
Lars Marowsky-Brée
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
_
bute.
#3 0x400696a5 in crm_abort (file=0x806986c "../../../linux-ha/crm/crmd/lrm.c",
function=0x806a09c "get_lrm_resource", line=771,
assert_condition=0x806a0e2 "class != NULL", do_fork=1) at utils.c:1240
Sincerely,
Lars Marowsky-Brée
--
Best Regards,
Huang
Lars Marowsky-Bree wrote:
On 2006-06-15T10:37:37, Huang Zhen <[EMAIL PROTECTED]> wrote:
No, it is unclear because a resource _container_ is treated as if it was
merely an option on a primitive. Which it is not.
I added the label like "belong to" to show the "contain&
he support to adding more than one primivate resources to
clone or master/slave.
As you noticed, this doesn't work. You can have a group contained within
a clone or an m/s though, instead of the primitive.
Sincerely,
Lars Marowsky-Brée
--
Best Regards,
Huang Zhen
Linux-HA, Linux
Huang Zhen wrote:
I think that there is no big difference between two screens and one
splited screen.
Current confusing is caused by the unclear labels like "Advance ID".
I have improved it and I attached the screen shot here.
Any comment is welcome!
I also added the support to a
of function 'unpack_resources'
../../../../linux-ha/lib/crm/pengine/status.c:106: warning: implicit
declaration of function 'unpack_status'
gmake[3]: *** [status.lo] Error 1
gmake[2]: *** [all-recursive] Error 1
gmake[1]: *** [all-recursive] Error 1
make: *** [all-recursive] Er
libs is not really relevant.
The GUI doesn't interpret the CIB itself.
It uses the pengine lib to get the information.
Well, the pengine lib analyses the CIB.
So you can say that the GUI interprets the CIB indirectly.
Sincerely,
Lars Marowsky-Brée
--
Best Regards,
Huang Zhen
Lin
>
> That clones are a pain to configure, and that the GUI simply doesn't
> allow several things which could be configured to be editted is probably
> simply a deficiency which will get fixed.
>
The attributes of primivate/group/clone/master can be edited in GUI for several
days.
I have tested t
ike that.
The GUI just know the pieces of XML which CRM can accepted as parameters.
The GUI never know the CIB. The CRM/CIB does.
Best Regards,
Huang Zhen
>
> Sincerely,
> Lars Marowsky-Br�e
>
> --
> High Availability & Clustering
> SUSE Labs, Research an
lease tell
me or report a bug.
If you feel that GUI is harder than write a XML (in the case you may be a xml
expert or have a
prefect xml editor), the GUI doesn't prevent you to use the tranditional way:
just write a whole xml and run it.
--
Best Regards,
Huang Zhen
Linux-HA, Linux
iding overall guidance is a project leader job.
Letting them silently run into the wall (because one got tired of the
discussion or it slipped one's mind) is not, if I may be so frank.
Sincerely,
Lars Marowsky-Brée
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China
a.org/ReportingProblems
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha
at, we're reasonably future proof?
OK. My apologies to Huang Zhen for not reading his code. [Today's a
holiday in the US]. Obviously he did exactly what I wanted him to.
I actually thought of something else to add a few minutes ago, but it's
potentially somewhat ris
--- Alan Robertson <[EMAIL PROTECTED]> wrote:
> Huang Zhen: Have you run this with CTS?
Not yet. will do soon.
>
> --
> Alan Robertson <[EMAIL PROTECTED]>
>
> "Openness is the foundation and preservative of friendship... Let me
> claim from you
Hi,
If we concern that the low-level changes may affect the release of 2.0.6, I can
stop check in and
keep the work locally.
What I have already committed are:
1. VIPArip.in, it is tested on SLES9 only.
2. Added cluster name, this should be ok.
3. Added weight to nodes (this may need more test,
?
Thanks,
Huang Zhen
--- Lars Marowsky-Bree <[EMAIL PROTECTED]> wrote:
> On 2006-05-28T09:38:40, Alan Robertson <[EMAIL PROTECTED]> wrote:
>
> > I agree about starting on 2.0.6. T thought about it earlier, but I had
> > a sudden, immediate, need to do planning fo
two integers to quorum plugin.
There are several ways, so I showed my suggestion and asked for your comments
here :)
Thanks!
Best Regards,
Huang Zhen
--- Lars Marowsky-Bree <[EMAIL PROTECTED]> wrote:
> On 2006-05-29T17:13:00, Huang Zhen <[EMAIL PROTECTED]> wrote:
>
> >
Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha
Hi,
The problem I met in x86-64 is:
The libdir was assigned in "configure", Line 503:
libdir='${exec_prefix}/lib'
So the condition was always false in "configure.in":
if test x"$libdir" = x; then
...
Then "needs_64_bit_libs" would never be called.
So the libdir at last was /usr/lib instead of /
other protocols versions.
Thanks Guy Coates for his very helpful scripts.
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
#!/bin/sh
#
# License:GN
gards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.or
ing issue directly.)
--
Best Regards,
Huang Zhen
Linux-HA, Linux Technology Center, China Systems & Technology Lab
China Development Labs, Beijing Tel: 86-10-82782244 Ext. 2845
Email: [EMAIL PROTECTED]
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.
nt.o] Error 1
Casting between "gpointer user_data" and "int msg_type = (int) user_data;"
etc. at rev. 1.37 (two days ago). Could you take a look, please.
Huang Zhen: Many thanks for fixing the bug at line 99 a few hours ago.
Much appreciated.
Now I've seem your model for
peration.
Then all changes would be sent to this copy.
When user wants to commit, send the xml in local cib copy to cib daemon.
We can reuse the current cib code. Two functions may be enough:
copy_cib_conn() and commit_cib_conn().
And your suggestions on this batch operations?
Thanks!
--
Best R
v: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
--
Best Regards,
Huang Zhen
Linux Technology Center
IBM China Development Lab, Beijing
Telno: (8610)82782244-2845
___
Linux-
redundent code/structure that you can simplify. Last time you mentioned
rewriting CCM, I am not sure if
it is a good idea since it involves lots of time and testing. Let me
know if I can help.
thanks
-Guochun
Huang Zhen wrote:
Hello,
With the development of heartbeat 2, CCM(membership) m
et me
know if I can help.
thanks
-Guochun
Huang Zhen wrote:
Hello,
With the development of heartbeat 2, CCM(membership) may need be
reorganized to fit some new requirement from CRM.
Like provide the whole node list of cluster whatever the node is
active or dead.
And we find we may need im
thoughts about how to improvment CCM?
--
Best Regards,
Huang Zhen
Linux Technology Center
IBM China Development Lab, Beijing
Telno: (8610)82782244-2845
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
_POPPED
origin=crm_timer_popped ]
Feb 23 01:07:17 BadNews: Feb 23 01:03:18 hadev3 crmd: [30061]: info:
mask(../../../linux-ha/crm/crmd/fsa.c:ghash_print_node): Welcome reply not
received from: hadev1 3
Feb 23 01:07:17 BadNews: Feb 23 01:03:18 h
d", XML_BOOLEAN_FALSE);
___
Linux-ha-cvs mailing list
Linux-ha-cvs@lists.linux-ha.org
http://lists.community.tummy.com/mailman/listinfo/linux-ha-cvs
--
Best Regards,
Huang Zhen
Linux Technology Center
IBM China Development Lab, Beijing
Thank you very much for your using and suggestions!
They are very helpful for us.
Joachim Banzhaf wrote:
Hi Huang Zhen,
Just tested the admin gui (from todays cvs). Nice work, so far.
Maybe this is of help for it to become even better:
I use something like cvs update -d; ConfigureMe
] on ocf::IPaddr::DcIPaddr for
client 4411, its parameters: is_managed=[1] ip=[127.0.0.10] timeout=[4]
te-target-rc=[7] lrm-is-probe=[true] crm_feature_set=[1.0.3] stayed in
operation list for 5200 ms (longer than 5000 ms)
We have seen this before. But does anyone see it recently?
I will report
Otherwise, the first few nodes will kill the others when they get the quorum
when running the StartOnebyOne test.
--
Best Regards,
Huang Zhen
Linux Technology Center
IBM China Development Lab, Beijing
Telno: (8610)82782244-2845
___
Linux-HA
l to all
nodes.
And there is a GUI tool for do some simple configuration when the cluster is
running.
Please refer to www.linux-ha.org/GuiGuide.
--
Best Regards,
Huang Zhen
Linux Technology Center
IBM China Development Lab, Beijing
Telno: (861
d to roll this out _now_. I think we
should, if we decide to give it a thumbs up, be able to roll this out
after a weekend of test cycles.
Please review your bugzillas and pending changes in your workspaces and
let me know what you think about this.
Alan, do you want to drive the release?
S
irectory when
migration.
Maybe in some case it should not be ignore even it's a directory.
Anyway, this shows that only the CVS page was missing.
I've create new CVS page in wiki as same as original one.
Huang Zhen wrote:
Hi, someone is asking about it.
--
Best Regards,
Huang
Hi, someone is asking about it.
--
Best Regards,
Huang Zhen
Linux Technology Center
IBM China Development Lab, Beijing
Telno: (8610)82782244-2845
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux
The reason of running management daemon as root is related PAM authorization.
Management daemon uses PAM to verify the user from remote (or local) node.
Now the related code only works as root.
Is it possible that a program running as non-root can call PAM authorization?
If yes, how? Thanks!
Ala
Hello,
The imap email has problem now. So I use this email account.
I have installed a debian and test the php script.
However, the RewriteRule of apache can't work.
And when I change the content in wiki, the content in www can't update. I
believe this is my
configuration problem.
One chnage
iewcvs.cgi/linux-ha/lib/mgmt/mgmt_tls_lib.c?r1=1.3&r2=1.4
gnutls_transport_ptr does exist on my system, but I'm not sure
about others.
Huang Zhen pointed out to me on IRC that I should be building
with gnutls 1.1, which does indeed solve the problem.
Shouldn't we test this
OCF_RESKEY_action
Raid1
* OCF_RESKEY_raidconf
* OCF_RESKEY_raiddev
ServeRAID
* OCF_RESKEY_serveraid
OCF_RESKEY_mergegroup
WAS
OCF_RESKEY_config
OCF_RESKEY_port
WinPopup
OCF_RESKEY_hostfile
Xinetd
* OCF_RESKEY_service
--
Best Regards,
Huang Zhen
Li
ée
--
Best Regards,
Huang Zhen
Linux Technology Center
IBM China Development Lab, Beijing
Telno: (8610)82782244-2845
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
-ha.org/GuiOpenIssues
1. should we integrete the management daemon with apphbd so we can start/stop
heartbeat?
2. how to deal with the different pam configure file formats in different OSs?
More issues are welcome too.
Thank you very much!
--
Best Regards,
Huang Zhen
Linux Technology Center
IBM
Alan Robertson wrote:
Huang Zhen wrote:
Hi, Rodrigo Henriquez,
Thank you for trying the new management daemon and GUI code.
Now there are several issues to run the management daemon and GUI client.
1. we need install following packages on the machine running the GUI
client
swig
Alan Robertson wrote:
Huang Zhen wrote:
5. we need add mgmtd to ha.cf as root.
"
...
apiauth mgmtd uid=root
crm 1
respawn root/usr/lib/heartbeat/mgmtd -v
Shouldn't this be a resource? You need a cluster IP, after all, as
there is no default
Hi, Rodrigo Henriquez,
Thank you for trying the new management daemon and GUI code.
Now there are several issues to run the management daemon and GUI client.
1. we need install following packages on the machine running the GUI client
swig
python(2.4) and python-devel
pyt
Guochun Shi wrote:
Alan Robertson wrote:
Huang Zhen wrote:
Hi,
May someone mention this before but I don't know. Just some thoughts.
In release 1, we only support 2 nodes, so the heartbeat connections
are symmetric.
But in release 2, we can support more than 2 nodes, then in many
eep a connection map or topologic
map of cluster.
This may help user to use even one cable to increase the redundance.
Is it feasible?
--
Best Regards,
Huang Zhen
Linux Technology Center
IBM China Development Lab, Beijing
Telno: (8610)82782244-2845
___
80 matches
Mail list logo