new release: backup_vmware3.12a - BUGFIX RELEASE

[TheBlob]

Hello to all, I've just released version 3.12a of the script.
This is a bugfix release and hopefully fixes the following problems,
identified by other users and by me:

3.12a ( 2009-29-07 )
- DONE fix the bug that causes VMs that had snapshot failures to not
to be powered up again after the fallback suspend
- DONE fix the bug in the enablevm function that does not allow to
enable more than one VM at a time if using VMware Server 1.0x
- DONE fix the bug in the disablevm function that does not allow to
enable more than one VM at a time if using VMware Server 1.0x
- DONE change the behavior of the script in case a VM does fallback
while snapshotting; in this case the snapshot operation will
report as successful
- DONE remove the old unused functions
- DONE fix the permissions to give to the vmuser user in the
INSTALL.TXT file

Please let me know if the issues are resolved for you too (I've tested
and they are, for me).
I've also fixed issue n. 6 from the issue list and hopefully also
issue 5 that, however, needs feedback to be closed.

Good evening and have a nice backup!

Michele

[Par_Tojkander]

Excellent, thank you for quick fixing! :)

I hate to bug you, but could you also upload the new version to the
repository (source)?

Best regards, Pär

[ronaldlw]

Awesome! I've downloaded it and am setting up to test now. Quick
question, is there a way to have the archive name made unique so that
they aren't overwritten each time? I'm assuming there must be a way
to do this since there is code to delete .tar files after they are a
certain age, which doesn't make sense if they are over-written each
time?

Thanks, this is a great script!

Ron

[ronaldlw]

i noticed when testing that this message came up:

2009-07-29 [16:06:24]cannot continue, background snapshots are still
enabled
please disable them by adding the following two lines

mainMem.partialLazySave = "FALSE"
mainMem.partialLazyRestore = "FALSE"
in the /etc/vmware/config file
and then restart all of your running VMs
or set the relevant option in the VMWare console
HOST -> Settings -> Priority -> "Take and restore snapshots in the
background"
and then restart all of your running VMs

The backups still run but is this really a problem?

Thanks,
Ron

[ronaldlw]

I think it would be good to have a variable or flag and/or a command
line switch to keep the script from doing a suspend of vms when the
snapshot fails. It would be very uncool to me if my vms pause for an
hour or more when backing them up.

Thanks,
Ron

[Michele Pensotti]

Hello, I forgot to upload to the source repository; now I've uploaded.

HOWEVER, since tonight all the backups have reported a non-existent failure in resuming, I'd suggest to not to release patches based on this still-buggy 3.12a, since I'll be releasing soon (today or tomorrow) 3.12b to fix this issue.

But in the meantime, if you like and have time, you may check the 3.12a and locate the problem, I still haven't checked.

Thank you again.

Michele

---

Da: Par_Tojkander <par.to...@gmail.com>
A: vmsbackup <vmsb...@googlegroups.com>
Inviato: Mercoledì 29 luglio 2009, 18:21:25
Oggetto: [vmsbackup] Re: new release: backup_vmware3.12a - BUGFIX RELEASE

[Michele Pensotti]

I suggest you create a request in the "issues" section of the site.

Thank you,

    Michele

---

Da: ronaldlw <goo...@rmwells.com>
A: vmsbackup <vmsb...@googlegroups.com>
Inviato: Mercoledì 29 luglio 2009, 18:54:21

Oggetto: [vmsbackup] Re: new release: backup_vmware3.12a - BUGFIX RELEASE

[Michele Pensotti]

Background snapshots are EVIL!
What happens with background snapshots is that the vmrun snapshot command immediately returns control to the shell, and the snapshot keeps going on in the background.
The problem is that my script does not know that the snapshot is still running and thus the tar command complains that the files changed while being read..

So I strongly reccomend you to disable the background snapshots.

Live long and prosper,

    Michele

---

Da: ronaldlw <goo...@rmwells.com>
A: vmsbackup <vmsb...@googlegroups.com>

Inviato: Mercoledì 29 luglio 2009, 23:17:41

Oggetto: [vmsbackup] Re: new release: backup_vmware3.12a - BUGFIX RELEASE

[Michele Pensotti]

Yes I too was thinking that yesterday while I was working on version 3.12a.
I'll add that to the to-do list if you add it to the "issues" section of the site, ok?

Thanks for your collaboration,

    Michele

---

Da: ronaldlw <goo...@rmwells.com>
A: vmsbackup <vmsb...@googlegroups.com>

Inviato: Mercoledì 29 luglio 2009, 23:21:58

Oggetto: [vmsbackup] Re: new release: backup_vmware3.12a - BUGFIX RELEASE

[ronaldlw]

Michele,

Ah, i see.. thanks for the info.. you might want to check the code to
make sure it actually still aborts when background snapshots are
enabled.

I wonder if you poll using the vmware-vim-cmd (like the suggestion for
getting away from the vmrun command) if it would report the actual
state?

Thanks!
Ron

On Jul 30, 3:00 am, Michele Pensotti <gcanavra...@yahoo.it> wrote:
> Background snapshots are EVIL!
> What happens with background snapshots is that the vmrun snapshot command immediately returns control to the shell, and the snapshot keeps going on in the background.
> The problem is that my script does not know that the snapshot is still running and thus the tar command complains that the files changed while being read.
>

[10nico]

On 30 Lug, 15:21, ronaldlw <goo...@rmwells.com> wrote:
> Michele,
>
> Ah, i see.. thanks for the info.. you might want to check the code to
> make sure it actually still aborts when background snapshots are
> enabled.

OOPS!
You are indeed right, the sanity checks did not work correctly and
allowed the backup to take place even in case the background snapshots
were still enabled.
I have now fixed this issue, and the fix will be included in the 3.12b
version.

>
> I wonder if you poll using the vmware-vim-cmd (like the suggestion for
> getting away from the vmrun command) if it would report the actual
> state?

No, I'm not polling right now; I still haven't found the time to test
to see if the polling actually works/is usable.
Also, since VMware Server 1.0.x hasn't got a good alternative to the
vmrun command, I'd have to create a substantial difference in the
logic, depending on the VMware Server version in use; one that polls,
for VMware 2.0.x and one that does not poll, for VMware 1.0.x

And by the way, the vmrun method is functional, if you create the
vmbackup user and give it the correct permissions, so why waste time
on a working piece of code, even if it's not the most elegant one?

I do not have time to spare, I'm barely keeping up with the bugs right
now!

I hope you understand my point of view. :-)

>
> Thanks!
> Ron
>

You're welcome! :-)

Michele

[ronaldlw]

I totally understand about keeping the vmrun function, i think i
misunderstood from comments in the code that you were wanting to move
away from using vmrun for some reason.. If I ever have some free time
myself I might rework the code to do only 2.x so that I can get a
better handle on what it's doing.

Great script!

Ron