romu's questions

've read tons of threads about this problem, but nothing seems to work for me.

I'm trying to expand the LVM volume of virtual machine located on our vSphere. This VM runs Debian 9. I already made the same thing on a CentOS7 without any trouble. So here I am.

In vSphere, the VM has a 100 GB storage. And we want to expand the current storage to fill that space.

root@build:/home/robot# df -h
Filesystem                  Size  Used Avail Use% Mounted on
udev                        3.9G     0  3.9G   0% /dev
tmpfs                       799M  8.8M  791M   2% /run
/dev/mapper/build--vg-root   51G   45G  3.7G  93% /
tmpfs                       4.0G     0  4.0G   0% /dev/shm
tmpfs                       5.0M     0  5.0M   0% /run/lock
tmpfs                       4.0G     0  4.0G   0% /sys/fs/cgroup
/dev/sda1                   236M   37M  188M  17% /boot
tmpfs                       799M     0  799M   0% /run/user/1000
tmpfs                       799M     0  799M   0% /run/user/1001

You can see the build--vg-root storage is almost full.

I already created the /dev/sda3 physical partition to hold the expanded storage...

root@build:/home/robot# fdisk -l
Disk /dev/sda: 100 GiB, 107374182400 bytes, 209715200 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x0aed3b2c

Device     Boot  Start       End   Sectors  Size Id Type
/dev/sda1  *      2048    499711    497664  243M 83 Linux
/dev/sda2       501758 125827071 125325314 59.8G  5 Extended
/dev/sda3       499712    501757      2046 1023K 8e Linux LVM
/dev/sda5       501760 125827071 125325312 59.8G 8e Linux LVM

...but when I run the pvcreate, I get an error...classical:

root@build:/home/robot# pvcreate /dev/sda3
  Device /dev/sda3 not found (or ignored by filtering).

There is no filter in the /etc/lvm/lvm.conf file, which is OOTB. I rebooted the VM several times, tried :

root@build:/home/robot# partprobe -s
/dev/sda: msdos partitions 1 3 2 <5>
/dev/mapper/build--vg-root: loop partitions 1
/dev/mapper/build--vg-swap_1: loop partitions 1

...but also:

root@build:/home/robot# cat /proc/partitions
major minor  #blocks  name

   8        0  104857600 sda
   8        1     248832 sda1
   8        3       1023 sda3
   8        5   62662656 sda5
  11        0    1048575 sr0
 254        0   54272000 dm-0
 254        1    8388608 dm-1

No luck...or skills. Any help would be great. Thanks.

I've spent many hours searching the web about how to shrink KVM virtual disk images, especially for Windows guests, with no luck.

All I've found is to zeroize the VM free space, defrag the virtual disk (from Windows), and then run qemu-img convert -c ... (-c flag to compress).

I've a Windows 7 VM, with a 100 GB virtual drive size. Initially, this VDD spent 40 GB on the host storage. Once it's zeroized, the VDD eats for real 100 GB on the host. And qemu-img -c ... creates a 91 GB, which is not at all what I expected.

On the modernie web site, we can download W7 VMs which are less than 10 GB, how is this possible? Is there a way to "really" compress the VM images?

Thanks to @dyasny, I made a small test with virt-sparsity. I cleaned up the W7 VM disk, disable hibernation, so the VDD only consumes 20 GB. Degraded the disk again, and ran again sdelete -z. Running virt-sparsity with the --compress flag gives a 80 GB virtual drive. Far from what I would have hoped.

EDIT-2016-02-16: "Refreshing" this question because the method to shrink a VM discussed here is very efficient but has a major drawback: it deletes all VM snapshots. If someone knows how to shrink a VM while preserving snapshots, feel free to share!

I'm facing a pretty annoying issue with Samba. Here is my setup:

• Windows Server 2003 + Active Directory (users authentication)
• NAS Server (Ubuntu 12.04 + PBIS Open Broker so this NAS is registered within the AD + Samba 3.6.3)
• Some AD Groups on the WS AD Server
• Some shares on the NAS for the AD groups

Extract of the /etc/samba/smb.conf:

[boxes]
comment = Development boxes
valid users = @"DOMAIN\dev"
writable = yes
path = /media/raid/boxes
write list = @"DOMAIN\dev"

This "boxes" share is allowed in both reading and writing for the DOMAIN\dev group. And it works...except for my own account.

With my own account, I can access the general shares (shared with the "domain users" group), but not the more specific shares (in this last case, I'm always asked with my credentials with a "access denied" message). I've to say it was not only the case, this has worked flawlessly for more than a year, and this issue appeared recently without any specific change I can remember.

I'm asking for help because I'm currently not good enough at Samba to investigate what happens. I went through the following log files in /var/log/samba:

• log.smbd: nothing really significant
• log.nmbd: nothing really significant
• log.#my PC name#: too old modification date
• log.#my PC IP#: too old modification date
• log.#my domain login#: file is empty

I also checked my account is not locked, I tried to change my password. This issues happens from both my own station (Fedora) or from any Windows PC.

So, how can I diagnose precisely what happens in order to fix this?

Thanks.

EDIT1: As advised by @sam_pan_mariusz, I checked the following command

id #my login#
uid=76547207(#my login#) gid=76546561(domain^users) groups=76546561(domain^users),76548718(#a group#)

Obviously, it's far from being correct because my own login is member of 3 groups and only 1 is listed here (I mean specific groups not "domain users").

I made another testing, with the same command, I've seen my boss was in the "dev" group and coding is not his job. So, on the AD server I removed him from the dev group and ran:

id #my boss login#
uid=76547181(#my boss login#) gid=76546561(domain^users) groups=76546561(domain^users),76547165(#group 1#),76548718(#group 2#),76548790(#group 3#),76547162(dev)

This means the groups and their contents are not well refreshed on the NAS.

EDIT2: /etc/nsswitch.conf file

passwd:         compat lsass
group:          compat lsass
shadow:         compat

hosts:          files dns
networks:       files

protocols:      db files
services:       db files
ethers:         db files
rpc:            db files

netgroup:       nis

No winbind here, only lsass. I've read somewhere PBis Open Broker and Winbind are not compatible.

EDIT3: Found a tricky workaround

On the nas, connected through SSH, I ran:

su - #my login#

And exited immediately, but this forced the groups membership to be refreshed...for my login. It's great, but I need this to be automatic and for all domain users and groups.

EDIT4: The above workaround has worked one time. Doesn't work anymore.

I'm setting up a virtualization server for my company based on Fedora 19 and KVM. Our goal is to being able to convert and run a bunch of VirtualBox VMs.

So I setup my F19 server, and created a br0 network bridge with bridge-utils. I first tried with VirtualBox, and no trouble, my testing VM (Ubuntu Server 10.04) works as it did on Windows.

Then, I converted the VM to qcow format and added this new VM to Virt-Manager. I can start it, log in, etc, everything works...except the network.

• ifconfig returns only the "lo" loopback adapter
• lspci shows well the "Ethernet Controller" of the correct type as specified in Virt-Manager

I've tried to change from bridge to NAT, change the type of network adapter in Virt-Manager, nothing works.

Any idea how to fix this?