[sheepdog-users] Testing snapshot-object-reclaim

[Valerio Pachera]

> I updated the branch for prevent the above combination. The problem
> still happens under heavy load. But I believe that a number of leaked
> objects would be smaller than previous case. Could you try it when you
> have time?

I doesn't seems good :-(
After removing all snapshots and vdi, there still are 1.3G.
It takes a huge amount of time for removing the snapshots (12 minutes)
on a 5g vdi.
When all snapshots are there, the used space is 15G.
After removing all snapshots but the vdi, the used space is 17G.

Sheepdog daemon version 0.8.0_150_gd3a1938

root at test006:~# dog cluster format -c 2:1
using backend plain store

root at test006:~# dog vdi list
  Name        Id    Size    Used  Shared    Creation time   VDI id  Copies  Tag
  test         0  5.0 GB  864 MB  0.0 MB 2014-04-08 09:15   7c2b25    2:1

root at test006:~# dog node info
Id      Size    Used    Avail   Use%
 0      217 GB  432 MB  217 GB    0%
 1      423 GB  432 MB  422 GB    0%
 2      467 GB  432 MB  467 GB    0%
Total   1.1 TB  1.3 GB  1.1 TB    0%
Total virtual image size        5.0 GB

root at test006:~# dog vdi snapshot test
root at test006:~# dog vdi snapshot test
root at test006:~# dog vdi snapshot test
root at test006:~# dog vdi snapshot test
root at test006:~# dog vdi snapshot test
root at test006:~# dog vdi snapshot test

root at test006:~# dog vdi list
  Name        Id    Size    Used  Shared    Creation time   VDI id  Copies  Tag
s test         1  5.0 GB  864 MB  0.0 MB 2014-04-08 09:15   7c2b25    2:1
s test         2  5.0 GB  4.5 GB  228 MB 2014-04-08 09:16   7c2b26    2:1
s test         3  5.0 GB   20 MB  4.7 GB 2014-04-08 09:19   7c2b27    2:1
s test         4  5.0 GB  4.5 GB  240 MB 2014-04-08 09:19   7c2b28    2:1
s test         5  5.0 GB   36 MB  4.7 GB 2014-04-08 09:21   7c2b29    2:1
s test         6  5.0 GB   32 MB  4.7 GB 2014-04-08 09:21   7c2b2a    2:1
  test         0  5.0 GB  0.0 MB  4.7 GB 2014-04-08 09:21   7c2b2b    2:1

root at test006:~# dog node info
Id      Size    Used    Avail   Use%
 0      220 GB  5.0 GB  215 GB    2%
 1      425 GB  5.0 GB  420 GB    1%
 2      470 GB  5.0 GB  465 GB    1%
Total   1.1 TB  15 GB   1.1 TB    1%
Total virtual image size        5.0 GB

root at test006:/mnt/sheep# time for n in $(seq 1 6); do dog vdi delete
-s $n test; done

real    12m0.505s
user    0m0.000s
sys     0m0.004s

root at test006:/mnt/sheep# dog vdi list
  Name        Id    Size    Used  Shared    Creation time   VDI id  Copies  Tag
  test         0  5.0 GB  0.0 MB  4.7 GB 2014-04-08 09:21   7c2b2b    2:1

root at test006:/mnt/sheep# dog node info
Id      Size    Used    Avail   Use%
 0      219 GB  7.1 GB  212 GB    3%
 1      424 GB  2.4 GB  422 GB    0%
 2      468 GB  7.1 GB  461 GB    1%
Total   1.1 TB  17 GB   1.1 TB    1%
Total virtual image size        5.0 GB

root at test006:/mnt/sheep# time dog vdi delete test
real    1m15.860s
user    0m0.000s
sys     0m0.000s

root at test006:/mnt/sheep# dog node info
Id      Size    Used    Avail   Use%
 0      217 GB  570 MB  217 GB    0%
 1      423 GB  162 MB  423 GB    0%
 2      467 GB  570 MB  467 GB    0%
Total   1.1 TB  1.3 GB  1.1 TB    0%
Total virtual image size        0.0 MB

Do you need logs and other info?