git.net

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

[Openstack] Adding new Hard disk to Compute Node


Maybe I should point out more clearly that there are several ways of  
providing disk space for your instances.

If you choose file based storage for your instances (e.g. ephemeral  
disks as qcow images), you'll need a lot of space in  
/var/lib/nova/instances as ephemeral storage. If you delete an  
instance its disk is also gone.

Then there's cinder that can provide persistant storage to your  
instances or for additional volumes to existing VMs. If you delete an  
instance its disk will not be deleted (if you choose so).
Cinder can be configured with different backends, e.g. LVM or Ceph.  
The short description in [1] scratches only the top of this but maybe  
this helps understanding the basics. For now you can ignore the HA  
references.

So in conclusion you'll need to make a choice (for now) how to provide  
disk space for your instances (ephemeral or persistant). You'll see  
"phys_disk" grow if you provide more space to /var/lib/nova/instances,  
e.g. we use Ceph as backend and have /var/lib/nova/instances mounted  
on shared storage which gives us 22 TB of space:

Final resource view: name=compute2 phys_ram=64395MB used_ram=68048MB  
phys_disk=22877GB used_disk=490GB

If you use cinder with LVM these statistics will differ, of course.

I hope this clears it up a little bit.

Regards

[1] https://docs.openstack.org/ha-guide/storage-ha-backend.html


Zitat von Bernd Bausch <berndbausch at gmail.com>:

> Your node uses logical volume /h020--vg-root/ as its root filesystem.
> This logical volume has a size of 370GB:
>
> # lsblk -o NAME,FSTYPE,SIZE,MOUNTPOINT,LABEL
> NAME                FSTYPE        SIZE MOUNTPOINT LABEL
> (...)
> â??â??sdk5              LVM2_member 371.5G
> *  â??â??h020--vg-root   ext4        370.6G /*
>   â??â??h020--vg-swap_1 swap          976M [SWAP]
>
> Now you created another physical volume, //dev/sdb1/, and added it to
> volume group /h020-vg/. This increases the size of the *volume group*,
> but not the size of the *logical volume*.
>
> If you want to provide more space to instances' ephemeral storage, you
> could:
>
>   * increase the size of root volume /h020--vg-root/ using the
>     /lvextend/ command, then increase the size of the filesystem on it.
>     I believe that this requires a reboot, since it's the root filesystem.
>
> or
>
>   * create another logical volume, e.g. lvcreate -L1000GB -n
>     lv-instances h020-vg for a 1000GB logical volume, and mount it under
>     //var/lib/nova/instances/: mount /dev/h020-vg/lv-instances
>     /var/lib/nova/instances
>     (before mounting, create a filesystem on /lv-instances/ and transfer
>     the data from //var/lib/nova/instances/ to the new filesystem. Also,
>     don't forget to persist the mount by adding it to //etc/fstab/)
>
> The second option is by far better, in my opinion, as you should
> separate operating system files from OpenStack data.
>
> You say that you are new to OpenStack. That's fine, but you seem to be
> lacking the fundamentals of Linux system management as well. You can't
> learn OpenStack without a certain level of Linux skills. At least learn
> about LVM (it's not that hard) and filesystems. You will also need to
> have networking fundamentals and Linux networking tools under your belt.
>
> Good luck!
>
> Bernd Bausch
>
> On 8/9/2018 2:30 AM, Jay See wrote:
>> Hai Eugen,
>>
>> Thanks for your suggestions and I went back to find more about adding
>> the new HD to VG. I think it was successful. (Logs are at the end of
>> the mail)
>>
>> Followed this link
>> - https://www.howtoforge.com/logical-volume-manager-how-can-i-extend-a-volume-group
>>
>> But still on the nova-compute logs it still shows wrong phys_disk
>> size. Even in the horizon it doesn't get updated with the new HD added
>> to compute node.
>>
>> 2018-08-08 19:22:56.671 3335 INFO nova.compute.resource_tracker
>> [req-14a2b7e2-7703-4a75-9014-180eb26876ff - - - - -] Final resource
>> view: name=h020 phys_ram=515767MB used_ram=512MB
>> *phys_disk=364GB *used_disk=0GB total_vcpus=40 used_vcpus=0 pci_stats=[]
>>
>> I understood they are not supposed to be mounted
>> on /var/lib/nova/instances so removed them now.
>>
>> Thanks
>> Jay.
>>
>>
>> root at h020:~# vgdisplay
>>   --- Volume group ---
>>   *VG Name               h020-vg*
>>   System ID
>>   Format                lvm2
>>   Metadata Areas        1
>>   Metadata Sequence No  3
>>   VG Access             read/write
>>   VG Status             resizable
>>   MAX LV                0
>>   Cur LV                2
>>   Open LV               2
>>   Max PV                0
>>   Cur PV                1
>>   Act PV                1
>>   VG Size               371.52 GiB
>>   PE Size               4.00 MiB
>>   Total PE              95109
>> *  Alloc PE / Size       95105 / 371.50 GiB*
>> *  Free  PE / Size       4 / 16.00 MiB*
>>   VG UUID               4EoW4w-x2cw-xDmC-XrrX-SXBG-RePM-XmWA2U
>>
>> root at h020:~# pvcreate */dev/sdb1*
>>   Physical volume "/dev/sdb1" successfully created
>> root at h020:~# pvdisplay
>>   --- Physical volume ---
>>   PV Name               /dev/sdk5
>>   VG Name               h020-vg
>>   PV Size               371.52 GiB / not usable 2.00 MiB
>>   Allocatable           yes
>>   PE Size               4.00 MiB
>>   Total PE              95109
>>   Free PE               4
>>   Allocated PE          95105
>>   PV UUID               BjGeac-TRkC-0gi8-GKX8-2Ivc-7awz-DTK2nR
>>
>>   "/dev/sdb1" is a new physical volume of "5.46 TiB"
>>   --- NEW Physical volume ---
>>   PV Name               /dev/sdb1
>>   VG Name
>>   PV Size               5.46 TiB
>>   Allocatable           NO
>>   PE Size               0
>>   Total PE              0
>>   Free PE               0
>>   Allocated PE          0
>>   PV UUID               CPp369-3MwJ-ic3I-Keh1-dJJY-Gcrc-CpC443
>>
>> root at h020:~# vgextend /dev/h020-vg /dev/sdb1
>>   Volume group "h020-vg" successfully extended
>> root at h020:~# vgdisplay
>>   --- Volume group ---
>>   VG Name               h020-vg
>>   System ID
>>   Format                lvm2
>>   Metadata Areas        2
>>   Metadata Sequence No  4
>>   VG Access             read/write
>>   VG Status             resizable
>>   MAX LV                0
>>   Cur LV                2
>>   Open LV               2
>>   Max PV                0
>>   Cur PV                2
>>   Act PV                2
>>   VG Size               5.82 TiB
>>   PE Size               4.00 MiB
>>   Total PE              1525900
>> *  Alloc PE / Size       95105 / 371.50 GiB*
>> *  Free  PE / Size       1430795 / 5.46 TiB*
>>   VG UUID               4EoW4w-x2cw-xDmC-XrrX-SXBG-RePM-XmWA2U
>>
>> root at h020:~# service nova-compute restart
>> root at h020:~# lsblk -o NAME,FSTYPE,SIZE,MOUNTPOINT,LABEL
>> NAME                FSTYPE        SIZE MOUNTPOINT LABEL
>> sda                               5.5T
>> â??â??sda1              vfat          500M            ESP
>> â??â??sda2              vfat          100M            DIAGS
>> â??â??sda3              vfat            2G            OS
>> sdb                               5.5T
>> â??â??sdb1              LVM2_member   5.5T
>> sdk                               372G
>> â??â??sdk1              ext2          487M /boot
>> â??â??sdk2                              1K
>> â??â??sdk5              LVM2_member 371.5G
>>   â??â??h020--vg-root   ext4        370.6G /
>>   â??â??h020--vg-swap_1 swap          976M [SWAP]
>> root at h020:~# pvscan
>>   PV /dev/sdk5   VG h020-vg         lvm2 [371.52 GiB / 16.00 MiB free]
>>   PV /dev/sdb1   VG h020-vg         lvm2 [5.46 TiB / 5.46 TiB free]
>>   Total: 2 [5.82 TiB] / in use: 2 [5.82 TiB] / in no VG: 0 [0   ]
>> root at h020:~# vgs
>>   VG      #PV #LV #SN Attr   VSize VFree
>>   h020-vg   2   2   0 wz--n- 5.82t 5.46t
>> root at h020:~# vi /var/log/nova/nova-compute.log
>> root at h020:~# 
>>
>>
>> On Wed, Aug 8, 2018 at 3:36 PM, Eugen Block <eblock at nde.ag
>> <mailto:eblock at nde.ag>> wrote:
>>
>>     Okay, I'm really not sure if I understand your setup correctly.
>>
>>         Server does not add them automatically, I tried to mount them.
>>         I tried they
>>         way they discussed in the page with /dev/sdb only. Other hard
>>         disks I have
>>         mounted them my self. Yes I can see them in lsblk output as below
>>
>>
>>     What do you mean with "tried with /dev/sdb"? I assume this is a
>>     fresh setup and Cinder didn't work yet, am I right?
>>     The new disks won't be added automatically to your cinder
>>     configuration, if that's what you expected. You'll have to create
>>     new physical volumes and then extend the existing VG to use new disks.
>>
>>         In Nova-Compute logs I can only see main hard disk shown in
>>         the the
>>         complete phys_disk, it was supposed to show more  phys_disk
>>         available
>>         atleast 5.8 TB if only /dev/sdb is added as per my understand
>>         (May be I am
>>         thinking it in the wrong way, I want increase my compute node
>>         disk size to
>>         launch more VMs)
>>
>>
>>     If you plan to use cinder volumes as disks for your instances, you
>>     don't need much space in /var/lib/nova/instances but more space
>>     available for cinder, so you'll need to grow the VG.
>>
>>     Regards
>>
>>
>>     Zitat von Jay See <jayachander.it at gmail.com
>>     <mailto:jayachander.it at gmail.com>>:
>>
>>         Hai,
>>
>>         Thanks for a quick response.
>>
>>         - what do you mean by "disks are not added"? Does the server
>>         recognize
>>         them? Do you see them in the output of "lsblk"?
>>         Server does not add them automatically, I tried to mount them.
>>         I tried they
>>         way they discussed in the page with /dev/sdb only. Other hard
>>         disks I have
>>         mounted them my self. Yes I can see them in lsblk output as below
>>         root at h020:~# lsblk -o NAME,FSTYPE,SIZE,MOUNTPOINT,LABEL
>>         NAME                                          FSTYPE        SIZE
>>         MOUNTPOINT                   LABEL
>>         sda                                                         5.5T
>>         â??â??sda1                                        vfat          500M
>>                           ESP
>>         â??â??sda2                                        vfat          100M
>>                           DIAGS
>>         â??â??sda3                                        vfat            2G
>>                           OS
>>         sdb                                                         5.5T
>>         â??â??sdb1                                                      5.5T
>>         â??â??cinder--volumes-cinder--volumes--pool_tmeta                84M
>>         â?? â??â??cinder--volumes-cinder--volumes--pool                   5.2T
>>         â??â??cinder--volumes-cinder--volumes--pool_tdata               5.2T
>>           â??â??cinder--volumes-cinder--volumes--pool                   5.2T
>>         sdc                                                         5.5T
>>         â??â??sdc1                                        xfs           5.5T
>>         sdd                                                         5.5T
>>         â??â??sdd1                                        xfs           5.5T
>>         /var/lib/nova/instances/sdd1
>>         sde                                                         5.5T
>>         â??â??sde1                                        xfs           5.5T
>>         /var/lib/nova/instances/sde1
>>         sdf                                                         5.5T
>>         â??â??sdf1                                        xfs           5.5T
>>         /var/lib/nova/instances/sdf1
>>         sdg                                                         5.5T
>>         â??â??sdg1                                        xfs           5.5T
>>         /var/lib/nova/instances/sdg1
>>         sdh                                                         5.5T
>>         â??â??sdh1                                        xfs           5.5T
>>         /var/lib/nova/instances/sdh1
>>         sdi                                                         5.5T
>>         â??â??sdi1                                        xfs           5.5T
>>         /var/lib/nova/instances/sdi1
>>         sdj                                                         5.5T
>>         â??â??sdj1                                        xfs           5.5T
>>         /var/lib/nova/instances/sdj1
>>         sdk                                                         372G
>>         â??â??sdk1                                        ext2         
>>         487M /boot
>>         â??â??sdk2                                                        1K
>>         â??â??sdk5                                        LVM2_member 371.5G
>>           â??â??h020--vg-root                             ext4        370.6G /
>>           â??â??h020--vg-swap_1                           swap         
>>         976M [SWAP]
>>
>>         - Do you already have existing physical volumes for cinder
>>         (assuming you
>>         deployed cinder with lvm as in the provided link)?
>>         Yes, I have tried one of the HD (/dev/sdb)
>>
>>         - If the system recognizes the new disks and you deployed
>>         cinder with lvm
>>         you can create a new physical volume and extend your existing
>>         volume group
>>         to have more space for cinder. Is this a failing step or
>>         someting else?
>>         System does not recognize the disks automatically, I have
>>         manually mounted
>>         them or added them to cinder.
>>
>>         In Nova-Compute logs I can only see main hard disk shown in
>>         the the
>>         complete phys_disk, it was supposed to show more  phys_disk
>>         available
>>         atleast 5.8 TB if only /dev/sdb is added as per my understand
>>         (May be I am
>>         thinking it in the wrong way, I want increase my compute node
>>         disk size to
>>         launch more VMs)
>>
>>         2018-08-08 11:58:41.722 34111 INFO nova.compute.resource_tracker
>>         [req-a180079f-d7c0-4430-9c14-314ac4d0832b - - - - -] F
>>         inal resource view: name=h020 phys_ram=515767MB used_ram=512MB
>>         *phys_disk=364GB* used_disk=0GB total_vcpus=
>>
>>         40 used_vcpus=0 pci_stats=[]
>>
>>         - Please describe more precisely what exactly you tried and
>>         what exactly
>>         fails.
>>         As explained in the previous point, I want to increase the 
>>         phys_disk size
>>         to use the compute node more efficiently. So to add the HD to
>>         compute node
>>         I am installing cinder on the compute node to add all the HDs.
>>
>>         I might be doing something wrong.
>>
>>         Thanks and Regards,
>>         Jayachander.
>>
>>         On Wed, Aug 8, 2018 at 11:24 AM, Eugen Block <eblock at nde.ag
>>         <mailto:eblock at nde.ag>> wrote:
>>
>>             Hi,
>>
>>             there are a couple of questions rising up:
>>
>>             - what do you mean by "disks are not added"? Does the
>>             server recognize
>>             them? Do you see them in the output of "lsblk"?
>>             - Do you already have existing physical volumes for cinder
>>             (assuming you
>>             deployed cinder with lvm as in the provided link)?
>>             - If the system recognizes the new disks and you deployed
>>             cinder with lvm
>>             you can create a new physical volume and extend your
>>             existing volume group
>>             to have more space for cinder. Is this a failing step or
>>             someting else?
>>             - Please describe more precisely what exactly you tried
>>             and what exactly
>>             fails.
>>
>>             The failing neutron-l3-agent shouldn't have to do anything
>>             with your disk
>>             layout, so it's probably something else.
>>
>>             Regards,
>>             Eugen
>>
>>
>>             Zitat von Jay See <jayachander.it at gmail.com
>>             <mailto:jayachander.it at gmail.com>>:
>>
>>             Hai,
>>
>>
>>                 I am installing Openstack Queens on Ubuntu Server.
>>
>>                 My server has extra hard disk(s) apart from main hard
>>                 disk where
>>                 OS(Ubuntu)
>>                 is running.
>>
>>                 (
>>                 https://docs.openstack.org/cinder/queens/install/cinder-stor
>>                  
>> <https://docs.openstack.org/cinder/queens/install/cinder-stor>
>>                 age-install-ubuntu.html
>>                 )
>>                 As suggested in cinder (above link), I have been
>>                 trying to add the new
>>                 hard
>>                 disk but the other hard disks are not getting added.
>>
>>                 Can anyone tell me , what am i missing to add these
>>                 hard disks?
>>
>>                 Other info : neutron-l3-agent on controller is not
>>                 running, is it related
>>                 to this issue ? I am thinking it is not related to
>>                 this issue.
>>
>>                 I am new to Openstack.
>>
>>                 ~ Jayachander.
>>                 --
>>                 P  *SAVE PAPER â?? Please do not print this e-mail
>>                 unless absolutely
>>                 necessary.*
>>
>>
>>
>>
>>
>>             _______________________________________________
>>             Mailing list:
>>             http://lists.openstack.org/cgi-bin/mailman/listinfo/openstac
>>             <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstac>
>>             k
>>             Post to     : openstack at lists.openstack.org
>>             <mailto:openstack at lists.openstack.org>
>>             Unsubscribe :
>>             http://lists.openstack.org/cgi-bin/mailman/listinfo/openstac
>>             <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstac>
>>             k
>>
>>
>>
>>
>>         --
>>         â??
>>         P  *SAVE PAPER â?? Please do not print this e-mail unless absolutely
>>         necessary.*
>>
>>
>>
>>
>>
>>
>>
>> --
>> â??
>> P  *SAVE PAPER â?? Please do not print this e-mail unless absolutely
>> necessary.*
>>
>>
>> _______________________________________________
>> Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>> Post to     : openstack at lists.openstack.org
>> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack