You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@cloudstack.apache.org by "Vash_X@gmx.de" <Va...@gmx.de> on 2022/12/08 18:54:03 UTC
Volume-size on physical storage differs hugely from configured size in CS management
Hi everyone,
today i found some strange things going on in my setup and i currently
lacking a idea where to start looking for debugging.
Running currently CS 4.17.0.1 i noticed that the actual on-disk-size on our
nfs-storage isn't matching the configured volume-size in CS Management.
We are using kvm only.
As of now i found multiple machines which are using a 50GB sized service
offering for creation of root volume, which are bigger then 100GB on the
actual storage.
Connecting into the VMs and checking the amount of storage the VM is
thinking it is using is even less; for example:
VM with 50GB thin provisioned root volume; shown correctly in mgmt and db
-> used ammount of storage according to OS: 21GB of 50GB -> size on
nfs-storage for the volume-file: 120GB
I checked the values inside the CS DB as well as the settings for the
serviceofferings and everything seems to be completely fine.
Also the nfs server settings are looking good.
Did anybody has a suggestion on where to look into? As of know it seems i
am waisting a lot of storage without any particular reason.
Regards.
Re: Volume-size on physical storage differs hugely from configured size in CS management
Posted by "Vash_X@gmx.de" <Va...@gmx.de>.
Hi Wei,
thanks for you shared information. As always really helpfull.
Regards
Am Mi., 14. Dez. 2022 um 20:59 Uhr schrieb Wei ZHOU <us...@gmail.com>:
> Exactly.
>
> You can use 'qemu-img convert' to recreate the image.
>
> If you use recent libvirt/qemu and guest os, you can use volume snapshot on
> running vms, the memory will be lost but image will not be touched. see
> https://github.com/apache/cloudstack/pull/5297
>
> -Wei
>
> On Thursday, 8 December 2022, Vash_X@gmx.de <Va...@gmx.de> wrote:
>
> > Hi Wei,
> >
> > indeed. I am using snapshots every now and then, but i delete them quiet
> > fast.
> > From your question i would assume that the vm-snapshots are 'expending'
> the
> > Root-volume of the VM and when deleting the snapshot the volume won't be
> > shrinked to the "original" size?
> >
> > Regards,
> > Chris
> >
> > Am Do., 8. Dez. 2022 um 22:31 Uhr schrieb Wei ZHOU <
> ustcweizhou@gmail.com
> > >:
> >
> > > Do you use vm snapshot or download volume of running vms ?
> > >
> > > -Wei
> > >
> > >
> > > On Thursday, 8 December 2022, Vash_X@gmx.de <Va...@gmx.de> wrote:
> > >
> > > > Hi everyone,
> > > >
> > > > today i found some strange things going on in my setup and i
> currently
> > > > lacking a idea where to start looking for debugging.
> > > >
> > > > Running currently CS 4.17.0.1 i noticed that the actual on-disk-size
> on
> > > our
> > > > nfs-storage isn't matching the configured volume-size in CS
> Management.
> > > > We are using kvm only.
> > > > As of now i found multiple machines which are using a 50GB sized
> > service
> > > > offering for creation of root volume, which are bigger then 100GB on
> > the
> > > > actual storage.
> > > >
> > > > Connecting into the VMs and checking the amount of storage the VM is
> > > > thinking it is using is even less; for example:
> > > >
> > > > VM with 50GB thin provisioned root volume; shown correctly in mgmt
> and
> > db
> > > > -> used ammount of storage according to OS: 21GB of 50GB -> size on
> > > > nfs-storage for the volume-file: 120GB
> > > >
> > > > I checked the values inside the CS DB as well as the settings for the
> > > > serviceofferings and everything seems to be completely fine.
> > > > Also the nfs server settings are looking good.
> > > >
> > > > Did anybody has a suggestion on where to look into? As of know it
> > seems i
> > > > am waisting a lot of storage without any particular reason.
> > > > Regards.
> > > >
> > >
> >
>
Volume-size on physical storage differs hugely from configured size in CS management
Posted by Wei ZHOU <us...@gmail.com>.
Exactly.
You can use 'qemu-img convert' to recreate the image.
If you use recent libvirt/qemu and guest os, you can use volume snapshot on
running vms, the memory will be lost but image will not be touched. see
https://github.com/apache/cloudstack/pull/5297
-Wei
On Thursday, 8 December 2022, Vash_X@gmx.de <Va...@gmx.de> wrote:
> Hi Wei,
>
> indeed. I am using snapshots every now and then, but i delete them quiet
> fast.
> From your question i would assume that the vm-snapshots are 'expending' the
> Root-volume of the VM and when deleting the snapshot the volume won't be
> shrinked to the "original" size?
>
> Regards,
> Chris
>
> Am Do., 8. Dez. 2022 um 22:31 Uhr schrieb Wei ZHOU <ustcweizhou@gmail.com
> >:
>
> > Do you use vm snapshot or download volume of running vms ?
> >
> > -Wei
> >
> >
> > On Thursday, 8 December 2022, Vash_X@gmx.de <Va...@gmx.de> wrote:
> >
> > > Hi everyone,
> > >
> > > today i found some strange things going on in my setup and i currently
> > > lacking a idea where to start looking for debugging.
> > >
> > > Running currently CS 4.17.0.1 i noticed that the actual on-disk-size on
> > our
> > > nfs-storage isn't matching the configured volume-size in CS Management.
> > > We are using kvm only.
> > > As of now i found multiple machines which are using a 50GB sized
> service
> > > offering for creation of root volume, which are bigger then 100GB on
> the
> > > actual storage.
> > >
> > > Connecting into the VMs and checking the amount of storage the VM is
> > > thinking it is using is even less; for example:
> > >
> > > VM with 50GB thin provisioned root volume; shown correctly in mgmt and
> db
> > > -> used ammount of storage according to OS: 21GB of 50GB -> size on
> > > nfs-storage for the volume-file: 120GB
> > >
> > > I checked the values inside the CS DB as well as the settings for the
> > > serviceofferings and everything seems to be completely fine.
> > > Also the nfs server settings are looking good.
> > >
> > > Did anybody has a suggestion on where to look into? As of know it
> seems i
> > > am waisting a lot of storage without any particular reason.
> > > Regards.
> > >
> >
>
Re: Volume-size on physical storage differs hugely from configured size in CS management
Posted by "Vash_X@gmx.de" <Va...@gmx.de>.
Hi Wei,
indeed. I am using snapshots every now and then, but i delete them quiet
fast.
From your question i would assume that the vm-snapshots are 'expending' the
Root-volume of the VM and when deleting the snapshot the volume won't be
shrinked to the "original" size?
Regards,
Chris
Am Do., 8. Dez. 2022 um 22:31 Uhr schrieb Wei ZHOU <us...@gmail.com>:
> Do you use vm snapshot or download volume of running vms ?
>
> -Wei
>
>
> On Thursday, 8 December 2022, Vash_X@gmx.de <Va...@gmx.de> wrote:
>
> > Hi everyone,
> >
> > today i found some strange things going on in my setup and i currently
> > lacking a idea where to start looking for debugging.
> >
> > Running currently CS 4.17.0.1 i noticed that the actual on-disk-size on
> our
> > nfs-storage isn't matching the configured volume-size in CS Management.
> > We are using kvm only.
> > As of now i found multiple machines which are using a 50GB sized service
> > offering for creation of root volume, which are bigger then 100GB on the
> > actual storage.
> >
> > Connecting into the VMs and checking the amount of storage the VM is
> > thinking it is using is even less; for example:
> >
> > VM with 50GB thin provisioned root volume; shown correctly in mgmt and db
> > -> used ammount of storage according to OS: 21GB of 50GB -> size on
> > nfs-storage for the volume-file: 120GB
> >
> > I checked the values inside the CS DB as well as the settings for the
> > serviceofferings and everything seems to be completely fine.
> > Also the nfs server settings are looking good.
> >
> > Did anybody has a suggestion on where to look into? As of know it seems i
> > am waisting a lot of storage without any particular reason.
> > Regards.
> >
>
Re: Volume-size on physical storage differs hugely from configured size in CS management
Posted by Wei ZHOU <us...@gmail.com>.
Do you use vm snapshot or download volume of running vms ?
-Wei
On Thursday, 8 December 2022, Vash_X@gmx.de <Va...@gmx.de> wrote:
> Hi everyone,
>
> today i found some strange things going on in my setup and i currently
> lacking a idea where to start looking for debugging.
>
> Running currently CS 4.17.0.1 i noticed that the actual on-disk-size on our
> nfs-storage isn't matching the configured volume-size in CS Management.
> We are using kvm only.
> As of now i found multiple machines which are using a 50GB sized service
> offering for creation of root volume, which are bigger then 100GB on the
> actual storage.
>
> Connecting into the VMs and checking the amount of storage the VM is
> thinking it is using is even less; for example:
>
> VM with 50GB thin provisioned root volume; shown correctly in mgmt and db
> -> used ammount of storage according to OS: 21GB of 50GB -> size on
> nfs-storage for the volume-file: 120GB
>
> I checked the values inside the CS DB as well as the settings for the
> serviceofferings and everything seems to be completely fine.
> Also the nfs server settings are looking good.
>
> Did anybody has a suggestion on where to look into? As of know it seems i
> am waisting a lot of storage without any particular reason.
> Regards.
>
Re: Volume-size on physical storage differs hugely from configured size in CS management
Posted by "Vash_X@gmx.de" <Va...@gmx.de>.
I forgot to mention:
Looking into the mangement server under 'storage/volumes/' i can see that
the Utilisation is way over 100%, worst i found is 237%. Without any alarm
and notification...
Am Do., 8. Dez. 2022 um 19:54 Uhr schrieb Vash_X@gmx.de <Va...@gmx.de>:
> Hi everyone,
>
> today i found some strange things going on in my setup and i currently
> lacking a idea where to start looking for debugging.
>
> Running currently CS 4.17.0.1 i noticed that the actual on-disk-size on
> our nfs-storage isn't matching the configured volume-size in CS Management.
> We are using kvm only.
> As of now i found multiple machines which are using a 50GB sized service
> offering for creation of root volume, which are bigger then 100GB on the
> actual storage.
>
> Connecting into the VMs and checking the amount of storage the VM is
> thinking it is using is even less; for example:
>
> VM with 50GB thin provisioned root volume; shown correctly in mgmt and db
> -> used ammount of storage according to OS: 21GB of 50GB -> size on
> nfs-storage for the volume-file: 120GB
>
> I checked the values inside the CS DB as well as the settings for the
> serviceofferings and everything seems to be completely fine.
> Also the nfs server settings are looking good.
>
> Did anybody has a suggestion on where to look into? As of know it seems i
> am waisting a lot of storage without any particular reason.
> Regards.
>