gmane.linux.redhat.cluster

Re: cman error in corosync

Hammad Siddiqi — 2013-06-18T15:32:33

Dear Crissie,

Thanks for the update. Will be waiting for the patch for this,

Regards,

Hammad Siddiqi


On Tue, Jun 18, 2013 at 8:22 PM, Christine Caulfield <ccaulfie< at >redhat.com>wrote:

Re: cman error in corosync

Christine Caulfield — 2013-06-18T15:22:39

Oh sorry.

The bug is simply an error in the way that modclusterd calls into cman - 
it passes an invalid parameter so cman prints the message.
The bug has no repercussions apart from that message as modclusterd has 
a fallback command to get the information it needs. TBH the bugzilla 
says nothing more interesting that that, apart from the patch needed to 
fix it :)

I don't have any information about the release of a fixed package, sorry.

Chrissie

On 18/06/13 15:45, Hammad Siddiqi wrote:

Re: cman error in corosync

Hammad Siddiqi — 2013-06-18T14:45:08

Also, I had tried visiting the bugzilla link but my account does not have
sufficient privileges to get detailed info for this bug. It would be really
great if you enlighten what the bug really is.

Thanks a lot for your prompt replies


Hammad Siddiqi


On Tue, Jun 18, 2013 at 6:38 PM, Hammad Siddiqi <hsiddiqi< at >gmail.com> wrote:

Re: cman error in corosync

Hammad Siddiqi — 2013-06-18T13:38:46

Thanks for sharing this info. Would you please let know when the update
will be available in public packages and which versions.

Regards,

Hammad Siddiqi


On Tue, Jun 18, 2013 at 1:23 PM, Christine Caulfield <ccaulfie< at >redhat.com>wrote:

Re: cman error in corosync

Christine Caulfield — 2013-06-18T08:23:02

That's a bug in modclusterd, not cman. It's been fixed upstream in ricci 
but I'm not sure if it's out in public packages yet:

https://bugzilla.redhat.com/show_bug.cgi?id=951470

Chrissie

Re: cman error in corosync

Christine Caulfield — 2013-06-18T08:12:31

I don't have the start of this thread in my email client. Which error?

Chrissie

Re: cman error in corosync

Fabio M. Di Nitto — 2013-06-18T08:16:56

https://www.redhat.com/archives/linux-cluster/2013-April/msg00009.html

Cheers
Fabio

Re: question on storage

Rick Stevens — 2013-06-18T01:25:48

On 06/17/2013 05:44 PM, taiyeb jadliwala issued this missive:

Assuming the fiber channel switch's mapping is flat (every port can see
every other port), you could query the HBA on one server to see if it
sees the WWNs of the other server. You'd probably need one of the tools
that come with your HBAs to do this (e.g. "scli" for QLogic cards).
----------------------------------------------------------------------
- Rick Stevens, Systems Engineer, AllDigital    ricks< at >alldigital.com -
- AIM/Skype: therps2        ICQ: 22643734            Yahoo: origrps2 -
-                                                                    -
-    Admitting you have a problem is the first step toward getting   -
-    medicated for it.      -- Jim Evarts (http://www.TopFive.com)   -
----------------------------------------------------------------------

question on storage

taiyeb jadliwala — 2013-06-18T00:44:22

Hello Experts,

I have a simple query.

Let say i am having two computer which are connected on fiber channel
switch. How can i check the connectivity between these two computer. For eg
if the same two computer are connected on ethernet i can use tcp/ip
protocol and i can ping command to test the connectivity similarly how can
i check the connectivity between two computers when connected using fiber
channel.

Re: cman error in corosync

Fabio M. Di Nitto — 2013-06-17T20:10:45

The fix should be upstream already.

Chrissie do you know if it's been included?

Fabio

Re: cman error in corosync

Hammad Siddiqi — 2013-06-17T19:05:00

Dear Fabio,

Thanks for the update. Is there any fix for this bug. Would really
appreciate if some patch or update is provided.

Thank you,

Hammad Siddiqi

Re: kernel BUG at fs/gfs2/glock.c

Ali Bendriss — 2013-06-17T16:15:43

Hello Ali Do you try to aske in kernel mailling list? Thanks 

Thanks for your email,
Just after sending to the list I've seen my mistake.
cluster 3.2 dlm library was running against the wrong linux kernel.
It seems that during the compilation cluster looks for the kernel in 
/usr/src/linux (and for me it was a symlink to an other linux version).

So I have cleaned /usr/src to link with the good kernel, recompiled the 
cluster package and all is fine now.

Re: kernel BUG at fs/gfs2/glock.c

emmanuel segura — 2013-06-17T15:42:56

Hello Ali

Do you try to aske in kernel mailling list?

Thanks


2013/6/12 Ali Bendriss <ali.bendriss< at >gmail.com>

kernel BUG at fs/gfs2/glock.c

Ali Bendriss — 2013-06-12T14:13:02

Hello,

I have enabled the quota on a gfs2 file system through fstab on a 2 (identical) 
nodes cluster : quota=on
but I've not used it yet.
Now I can see the kernel message in attachment only on one node.

thanks


sotware version:
linux  3.9.3
cluster 3.2.0
gfs2-utils 3.1.5

--
AliJun 12 04:47:37 minnie kernel: [34757.502162] original: do_sync+0x189/0x4c0 [gfs2]
Jun 12 04:47:37 minnie kernel: [34757.502167] pid: 10880
Jun 12 04:47:37 minnie kernel: [34757.502169] lock type: 8 req lock state : 1
Jun 12 04:47:37 minnie kernel: [34757.502176] new: do_sync+0x189/0x4c0 [gfs2]
Jun 12 04:47:37 minnie kernel: [34757.502177] pid: 10880
Jun 12 04:47:37 minnie kernel: [34757.502179] lock type: 8 req lock state : 1
Jun 12 04:47:37 minnie kernel: [34757.502184]  G:  s:EX n:8/0 f:Iqb t:EX d:EX/0 a:0 v:0 r:6 m:200
Jun 12 04:47:37 minnie kernel: [34757.502192]   H: s:EX f:cH e:0 p:10880 [gfs2_quotad] do_sync+0x189/0x4c0 [gfs2]
Jun 12 04:47:37 minnie kernel: [34757.502221] ------------[ cut here ]------------
Jun 12 04:47:3

fenced_domain_info error -1

urgrue — 2013-06-06T18:17:45

I seem to get rgmanager stuck/unresponsive/unkillable rather regularly 
after a fence attempt has failed. fence_ack_manual rarely brings it back 
into shape and I have to resort to rebooting the node(s).
Googling the error messages usually gets me nowhere, e.g. the one below 
(fenced_domain_info error -1) yields exactly one hit, the source code. I 
feel special.

Can anyone point me in the right direction on how to debug this further?

This is RHEL 6.3 and a simple 2-node cluster + quorum disk and shared 
SAN storage with HA-LVM.

Here's a recent log snippet if it helps:
Jun 06 10:07:23 dlm_controld cluster node 1 removed seq 100812
Jun 06 10:07:23 dlm_controld del_configfs_node rmdir 
"/sys/kernel/config/dlm/cluster/comms/1"
Jun 06 10:07:23 dlm_controld dlm:controld conf 1 0 1 memb 2 join left 1
Jun 06 10:07:23 dlm_controld dlm:ls:rgmanager conf 1 0 1 memb 2 join left 1
Jun 06 10:07:23 dlm_controld rgmanager add_change cg 4 remove nodeid 1 
reason 3
Jun 06 10:07:23 dlm_controld rgmanager add_change cg 4 cou

Using LVM / ext4 in cluster.conf more than one time

Nicolas Ross — 2013-06-06T16:37:06

Hi !

We have 2 clusters of 8 nodes each. Since we begun using RHCS about 2.5
years ago, we mostly use GFS on a shared storage array for data storage.
In most cases, there is no need to use GFS, ext4 would be enough since
that filesystem is only used within one service.

Production services are enabled on one cluster at the time. A service
containing a webserver and data directories is running in cluster A. A
service also exists in cluster B that only use the data directories.
Data is then synced from cluster A to cluster B manually for recovery in
case of disaster. So in each cluster, we have a copy of 2 services. Both
use the same data file system, but only one start the webserver. I will
upload a verry simplified version of one cluster.conf to illustrate this
and the problem.

We begun migrating some GFS file system to ext4, as it offers more
performance, while retaining the same HA features in the event of a node
failure. And while doing so, we discover this problem.

While a ext3 using HA-LVM in a clust

Re: fence_ovh - Fence agent for OVH

Adrian Gibanel — 2013-06-05T14:52:55

----- Mensaje original ----- 




Ok. 

I think it's just a matter of trying it. Not sure when I will have the time.



Well, I didn't find a default field for setting it and I added one of my own. I think that you can use OVH API without that email informed but I prefer myself to be informed about fencing. So, I think that rather than writing in a file somewhere in the filesystem, it would be more useful to write it in cluster.conf.

  I don't think if I understood what you meant. You were thinking that maybe email should be a cluster wide settings instead of a fence agent wide setting... but then you thought that the latter was better idea?

Ok. I'll contribute back to it if I can when it's public.

Thank you!

Fence agent howto needed

Adrian Gibanel — 2013-06-02T21:26:43

As I mentioned in a previous message I think there's missing a: 

How to write your own Fence agent howto 
or 
Fence agent howto 

As I've written one fence agent myself I think I can help in explaining how to use included python fencing libraries (when you program in python of course) which eases a lot the parsing of options from stdin. I mean, it's easy, but it's far from obvious even by checking the default python fence_* scripts. 

Thank you.

fence_ovh - Fence agent for OVH

Adrian Gibanel — 2013-06-02T21:12:54

  As requested by digimer in linux-ha irc channel here there is fence_ovh. It's not a priority that it's included by default in official distribution of cluster software but if you guide me on how to polish it I think I can improve it a lot more and make tests in real machines (as long as my machines are still test machines and not production ones). 

1) What is fence_ovh 

fence_ovh is a fence agent based on python for the big French datacentre provider OVH. You can get information about OVH on: http://www.ovh.co.uk/ . I also wanted to make clear that I'm not part of official OVH staff. 

2) Features 
The script has two main functions: 

* Reboot into rescue mode (action=off) 
* Reboot into the hard disk (action=on;action=reboot) 

3) Technical details 
So as you might deduce the classical fence mechanism which turns off the other node is not actually done by turning off the machine but by rebooting it into a rescue mode. 

Another particular thing to mention is that the script checks if the machine has reb

Re: FenceAgentAPI needs to be updated

Digimer — 2013-05-31T13:49:25

If you can give me the rough notes/point-form list on what is required, 
I will write it out in long form for the wiki. Also, it's been long 
enough that it would be good, if you have the time, to review the 
document for any other missing or deprecated bits. Then I can give the 
whole thing a once-over at the same time.

Re: FenceAgentAPI needs to be updated

Marek Grac — 2013-05-31T07:14:31

Hi,

I believe that I can be that source, feel free to mail me questions.

m,