issues with snapshot in R77.30

[cciesec2006]

Hi folks,

something wrong with snapshot utility? I have a Dell R720 running R77.30 with HFA_205 as a stand-alone log server.

When I run the command "show snapshots", it supposes to show me all the snaphots that I've been taking so far but it does NOT. See below

[Expert@fwlog:0]# installed_jumbo_take
R77.30 Jumbo Hotfix Accumulator take_205 is installed, see sk106162.
[Expert@fwlog:0]#
fwlog> show snapshots
Creation of an additional restore point will need 12.768G
Amount of space available for restore points is 2532.91G
fwlog>
fwlog> show snapshot

fwlog_HFA_205 fwlog_HFA_185 fwlog_HFA_178 fwlog_HFA_159
fwlog> show snapshot fwlog_HFA_

Has anyone seen this before? New bugs :-(

thanks in advance

[Rls56]

Good Afternoon,

There is something wrong with the snapshot utility? I have an R77.30 VMware cluster with Take 204.

When using the WebUI Maintenance Utility to create a snapshot, it shows under the Snapshot Management window that the available space is FREE;
but under STATISTICS it show's "Creation of an additional image will require 6.240G
amount of space available for images is 0.02G.

df -h show that I have 100G available
Also export is greyed out.

[Bob_Zimmerman]

The space Gaia uses for snapshots doesn't show up in 'df'. I'm betting somebody has handed all of the space on the disk to lv_current, leaving no space for snapshots. You can check this with the command 'lvm_manager'. It should show a large amount of unallocated space, but I bet on your system it doesn't.

For a more technical discussion, Gaia uses a Linux tool called LVM for managing "partitions" (actually volumes) and for taking snapshots. The disks in a box are all given to an LVM volume group named vg_splat:

[Expert@DallasSA]# pvscan
  PV /dev/sda3   VG vg_splat        lvm2 [356.69 GiB / 255.69 GiB free]
  Total: 1 [356.69 GiB] / in use: 1 [356.69 GiB] / in no VG: 0 [0   ]

[Expert@DallasSA]# vgs
  VG       #PV #LV #SN Attr   VSize   VFree
  vg_splat   1   4   0 wz--n- 356.69g 255.69g

[Expert@DallasSA]# vgdisplay
  --- Volume group ---
  VG Name               vg_splat
  System ID
  Format                lvm2
  Metadata Areas        1
  Metadata Sequence No  15
  VG Access             read/write
  VG Status             resizable
  MAX LV                0
  Cur LV                4
  Open LV               2
  Max PV                0
  Cur PV                1
  Act PV                1
  VG Size               356.69 GiB
  PE Size               32.00 MiB
  Total PE              11414
  Alloc PE / Size       3232 / 101.00 GiB
  Free  PE / Size       8182 / 255.69 GiB
  VG UUID               55Qy6m-0Euw-7OlF-X33e-SIjB-SiB5-XHGAtx

[Expert@DallasSA]# lvs
  LV          VG       Attr       LSize  Pool Origin Data%  Meta%  Move Log Cpy%Sync Convert
  hwdiag      vg_splat -wi-a-----  1.00g
  lv_current  vg_splat -wi-ao---- 32.00g
  lv_fcd_GAIA vg_splat -wi-a-----  8.00g
  lv_log      vg_splat -wi-ao---- 60.00g

This machine has a 400 GB SSD and is running R81. I think the volume sizes are all default, but I may have increased the space given to lv_current.

When you take a snapshot, here's what happens:

[Expert@DallasSA]# clish

DallasSA> add snapshot TestSnap
Taking snapshot. You can continue working normally.
You can use the command 'show snapshots' to monitor creation progress.

DallasSA> exit

[Expert@DallasSA]# lvs
  LV              VG       Attr       LSize  Pool Origin     Data%  Meta%  Move Log Cpy%Sync Convert
  hwdiag          vg_splat -wi-a-----  1.00g
  lv_TestSnap     vg_splat -wi-ao---- 17.31g
  lv_current      vg_splat owi-aos--- 32.00g
  lv_current_snap vg_splat swi-aos--- 29.00g      lv_current 0.01
  lv_fcd_GAIA     vg_splat -wi-a-----  8.00g
  lv_log          vg_splat -wi-ao---- 60.00g

[Expert@DallasSA]# vgs
  VG       #PV #LV #SN Attr   VSize   VFree
  vg_splat   1   6   1 wz--n- 356.69g 209.38g

### Some time later... ###

[Expert@DallasSA]# lvs
  LV          VG       Attr       LSize  Pool Origin Data%  Meta%  Move Log Cpy%Sync Convert
  hwdiag      vg_splat -wi-a-----  1.00g
  lv_TestSnap vg_splat -wi-a----- 17.31g
  lv_current  vg_splat -wi-ao---- 32.00g
  lv_fcd_GAIA vg_splat -wi-a-----  8.00g
  lv_log      vg_splat -wi-ao---- 60.00g

[Expert@DallasSA]# vgs
  VG       #PV #LV #SN Attr   VSize   VFree
  vg_splat   1   5   0 wz--n- 356.69g 238.38g

It makes two new logical volumes. lv_current_snap is a snapshot of lv_current, then lv_TestSnap. Note that the VFree number in lv_splat drops from 255.69 GB earlier to 209.38 GB, a decrease of 46.31 GB. That's because it allocated 29 GB to lv_current_snap and 17.31 GB to lv_TestSnap. If I had less than 46 GB available, it wouldn't be able to take a snapshot, even though the end state of the snapshot only consumes 17.31 GB when it's done.