[drbd-mc] Announce: LCMC 1.3.0 / Pacemaker, DRBD, KVM GUI

[Rasto Levrinc]

Hi,

this is the next release of the Linux Cluster Management Console, the
high availability GUI. This release brings mostly usability
improvements.

[Caspar Smit]

Hi Rasto,

I checked the new LCMC version today and found out some stuff again.

1) I cannot right-click on resources/groups in the graphical display
anymore to get the sub-menu. Not even on the hosts.
It is only possible to right click on the resource/group names in
the left window (under CRM pacemaker)
Is this a 'feature' or a 'bug' ?

2) Everytime I connect to a cluster I get a big red <cluster name>:
Cluster status failed scrolling but after that the cluster seems fine.
This is when using sudo and root.

3) disconnecting from a cluster doesn't really disconnect because when
I disconnect from a cluster and connect another cluster it still does
some status updates for the disconnected cluster (I get an: updating
drbd/cluster status from the disconnected cluster(s), and that some
clusters don't even have drbd installed and this takes a lot of time).
how can i really disconnect the cluster(s) without leaving LCMC?

4) I found out today I had a split-brain cluster and I noticed that
the pacemaker Master/slave resource had no problem and "all is fine"
Primary/Secondary etc... I tried resolving this through LCMC (DRBD)
but all my options were greyed out for doing anything to the DRBD
resources. Using sudo and root didn't make a difference. Any idea why?

5) Can you explain the "workaround for iSCSITarget RA" ? I'm using a
cluster with that RA but nothing seems changed.

6) What does the "disable one-host cluster by default" exactly do
because I can still add 1 host, and then a cluster with that 1 host
(and this is something I would still like to able to do if
nessessary).

Kind regards,

Caspar Smit


Op 26 februari 2012 22:39 heeft Rasto Levrinc

Post by Rasto Levrinc
Hi,
this is the next release of the Linux Cluster Management Console, the
high availability GUI. This release brings mostly usability
improvements.
From now on it is possible to operate on multiple resources at once,
thus eliminating some of the repetitive work. You can select services,
block devices and hosts by drawing a rectangle around them, or
ctrl-click them. The menus in this context work on all selected
components. For example you can select multiple services and migrate
them all at once, instead of doing it one by one.
http://sourceforge.net/apps/gallery/lcmc/index.php?g2_itemId=40
As you probably don't know there were options to connect to the existing
cluster quickly, but since kids these days don't even know what a
command line option is, there is now an in-your-face button for that.
Much work went into reducing the number of SSH connection and sudo calls
in this release as well, because they used to log a lot.
http://lcmc.sf.net
http://sourceforge.net/apps/gallery/lcmc/
https://github.com/rasto/lcmc
* implement quick-connect-to-the-cluster box
* implement a ignore-defaults resource agent list
* workaround for iSCSITarget RA
* enable sudo checkbox if user is not root automatically
* disable one-host cluster by default
* add --one-host-cluster option
* rewrite code that collects info from the cluster
* fixes a null pointer exception while clicking on the host
* implement multi selection
Rasto Levrinc
_______________________________________________
drbd-mc mailing list
drbd-mc at lists.linbit.com
http://lists.linbit.com/mailman/listinfo/drbd-mc

[Rasto Levrinc]

Post by Caspar Smit
Hi Rasto,
I checked the new LCMC version today and found out some stuff again.
1) I cannot right-click on resources/groups in the graphical display
anymore to get the sub-menu. Not even on the hosts.
? It is only possible to right click on the resource/group names in
the left window (under CRM pacemaker)
? Is this a 'feature' or a 'bug' ?

Hi Caspar,

it's a bug that effects Windows users, so I got a lot of reports about
that. I've just fixed that, so later today there'll be 1.3.1 version.

Post by Caspar Smit
Cluster status failed scrolling but after that the cluster seems fine.
This is when using sudo and root.

Yup, will be fixed, it's a cosmetic problem.

Post by Caspar Smit
3) disconnecting from a cluster doesn't really disconnect because when
I disconnect from a cluster and connect another cluster it still does
some status updates for the disconnected cluster (I get an: updating
drbd/cluster status from the disconnected cluster(s), and that some
clusters don't even have drbd installed and this takes a lot of time).
how can i really disconnect the cluster(s) without leaving LCMC?

This is because of the effort to reduce the number ssh connections. At
the moment you can only kill these processes on the server, until I fix
it.

Post by Caspar Smit
4) I found out today I had a split-brain cluster and I noticed that
the pacemaker Master/slave resource had no problem and "all is fine"
Primary/Secondary etc... I tried resolving this through LCMC (DRBD)
but all my options were greyed out for doing anything to the DRBD
resources. Using sudo and root didn't make a difference. Any idea why?

Pacemaker can't do anything in this case with the resource. You should
inspect why you had the split-brain, maybe you're missing some
configuration.

To find out if you have split-brain from the GUI, you have to connect
the disconnected resource, it will not connect but it will detect that
it has a split brain. Connecting is enabled (if it is used by Pacemaker)
only in advanced mode. The GUI gets message that there's a split-brain
and offers the split-brain resolution dialog.

Post by Caspar Smit
5) Can you explain the "workaround for iSCSITarget RA" ? I'm using a
cluster with that RA but nothing seems changed.

It shouldn't effect you at all. There is some remote possibility that
default for "iSCSI Target implementation" could be wrong. There was also
a "status " with space in the meta-data, that could get stored in the
CIB. It would be ignored. You should check if you have it and remove it.

Post by Caspar Smit
6) What does the "disable one-host cluster by default" exactly do
because I can still add 1 host, and then a cluster with that 1 host
(and this is something I would still like to able to do if
nessessary).

there's a --one-host-cluster option for the old functionality now. It
was disabled by default because someone could define one host cluster by
mistake and destroy the existing two-node DRBD configuration.

thanks,
Rasto

[Caspar Smit]

Rasto,

Thank you very much for your explanations, it cleared a lot.

I didn't know the split brain resolution was only available in
advanced mode, thanks now it works :)

Kind regards,

Caspar Smit

Op 28 februari 2012 21:02 heeft Rasto Levrinc

Post by Rasto Levrinc

Post by Caspar Smit
Hi Rasto,
I checked the new LCMC version today and found out some stuff again.
1) I cannot right-click on resources/groups in the graphical display
anymore to get the sub-menu. Not even on the hosts.
? It is only possible to right click on the resource/group names in
the left window (under CRM pacemaker)
? Is this a 'feature' or a 'bug' ?

Hi Caspar,
it's a bug that effects Windows users, so I got a lot of reports about
that. I've just fixed that, so later today there'll be 1.3.1 version.

Post by Caspar Smit
Cluster status failed scrolling but after that the cluster seems fine.
This is when using sudo and root.

Yup, will be fixed, it's a cosmetic problem.

Post by Caspar Smit
3) disconnecting from a cluster doesn't really disconnect because when
I disconnect from a cluster and connect another cluster it still does
some status updates for the disconnected cluster (I get an: updating
drbd/cluster status from the disconnected cluster(s), and that some
clusters don't even have drbd installed and this takes a lot of time).
how can i really disconnect the cluster(s) without leaving LCMC?

This is because of the effort to reduce the number ssh connections. At
the moment you can only kill these processes on the server, until I fix
it.

Post by Caspar Smit
4) I found out today I had a split-brain cluster and I noticed that
the pacemaker Master/slave resource had no problem and "all is fine"
Primary/Secondary etc... I tried resolving this through LCMC (DRBD)
but all my options were greyed out for doing anything to the DRBD
resources. Using sudo and root didn't make a difference. Any idea why?

Pacemaker can't do anything in this case with the resource. You should
inspect why you had the split-brain, maybe you're missing some
configuration.
To find out if you have split-brain from the GUI, you have to connect
the disconnected resource, it will not connect but it will detect that
it has a split brain. Connecting is enabled (if it is used by Pacemaker)
only in advanced mode. The GUI gets message that there's a split-brain
and offers the split-brain resolution dialog.

Post by Caspar Smit
5) Can you explain the "workaround for iSCSITarget RA" ? I'm using a
cluster with that RA but nothing seems changed.

It shouldn't effect you at all. There is some remote possibility that
default for "iSCSI Target implementation" could be wrong. There was also
a "status " with space in the meta-data, that could get stored in the
CIB. It would be ignored. You should check if you have it and remove it.

Post by Caspar Smit
6) What does the "disable one-host cluster by default" exactly do
because I can still add 1 host, and then a cluster with that 1 host
(and this is something I would still like to able to do if
nessessary).

there's a --one-host-cluster option for the old functionality now. It
was disabled by default because someone could define one host cluster by
mistake and destroy the existing two-node DRBD configuration.
thanks,
Rasto
--
Dipl.-Ing. Rastislav Levrinc
rasto.levrinc at gmail.com
Linux Cluster Management Console
http://lcmc.sf.net/
_______________________________________________
drbd-mc mailing list
drbd-mc at lists.linbit.com
http://lists.linbit.com/mailman/listinfo/drbd-mc