Re: Node id uniqueness is not checked

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Dell - Internal Use - Confidential

Hello,

 

Background:

We are trying to introduce clustering to our embedded environment

Our environment includes following packages and modules for the initial test:

 

1. Pacemaker - version 1.1.10

2. Corosync - version 2.3.4

3. Corosync.conf file find it attached to this email thread

4. 2 Nodes in the cluster (master and slave)

5. Test app1 that publishes some sensor data (only Master node publishes, slave node just receives the data)

6. Test app2 that receives the published data and outputs on the screen (only master node outputs the data on the screen)

 

Our test has been successful and we are able to create a cluster with 2 nodes and everything seems to be working fine. 

However, we are observed that corosync is consuming about 100MB of RAM when both the test applications are alive.  100M is a lot memory…

 

Problem:

We would like to know if there are any configuration settings or tuning that we need to perform to reduce the memory usage. It is very critical to reduce the memory consumption as we are running it in embedded environment.

 

We have talked with the pacemaker mailing list.  We have tried

1.       Reduce the PCMK_ipc_buffer

2.      Change the PCMK_ipc_buffer from shared memory to socket memory.  

 

The memory for corosync is still consuming 100MB. More data below….

 

Data:

 

Virt

Res

Shr

PCMK_ipc_buffer

PCMK_ipc_type

Corosync

100m

99m

69m

5128

Socket

corosync

100m

97m

69m

5128

share

 

Process

Virt

res

Attrd

21308

2744

Cib

29608

9676

Crmd

72932

5152

Pacemaker

31380

2636

Stonithd

23072

4028

lrmd

10448

3012

 

 

Df /dev/shm with the PCMK_ipc_type set to socket printed the following.

 

Filesystem     1K-blocks  Used Available Use% Mounted on

tmpfs             124812 33492     91320  27% /dev/shm

 

The IPC portion is using about 27% of the memory aka 33.5m.

 

 

Thanks,
Heidi Maeder


-----Original Message-----
From: discuss-bounces@xxxxxxxxxxxx [mailto:discuss-bounces@xxxxxxxxxxxx] On Behalf Of Christine Caulfield
Sent: Thursday, January 29, 2015 3:37 AM
To: discuss@xxxxxxxxxxxx
Subject: Re: Node id uniqueness is not checked

On 28/01/15 07:56, Vladislav Bogdanov wrote:
> Hi!
>
> Just discovered that corosync doesn't check uniqueness of the node id
> in the config file (did't check is that true for online udpu member add).
>
> That leads to an abort on a node listed later in config. Other node
> becomes unresponsive (I tested in two-node cluster, so cannot say
> about other nodes in bigger cluster).
>
> It would be nice to add a check so 'corosync -t' complains.
>

That's a really good suggestion, I'll add it to my 'To Do' list. Thank you.

Chrissie

_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss

_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss

[Index of Archives]     [Linux Clusters]     [Corosync Project]     [Linux USB Devel]     [Linux Audio Users]     [Photo]     [Yosemite News]    [Yosemite Photos]    [Linux Kernel]     [Linux SCSI]     [X.Org]

  Powered by Linux