tests: Add a pre-allocation test #284
Conversation
|
Note that you can also point this test to a build directory ( |
EchterAgo
force-pushed
the
preallocation_test
branch
2 times, most recently
from
aa6aaef to
e3b329c
Compare
|
Looks like checkstyle fails because we do not have openzfs@9192ab7 |
|
See openzfs#15284 |
EchterAgo
force-pushed
the
preallocation_test
branch
from
e3b329c to
567a9d2
Compare
|
Rebased to have the checkstyle fix. |
EchterAgo
force-pushed
the
preallocation_test
branch
from
567a9d2 to
74e9d59
Compare
|
Ahh dang. This actually looks like it might have useful things. |
|
@andrewc12 yea, we should work on building a bit of a test library with common functionality. I wanted to also go over Is there any reason the tests spend so much time sleeping? I'm currently trying to reduce that too. |
|
Mostly extra cautious to avoid BSODs |
|
I need to fix the Python style of this change. |
EchterAgo
force-pushed
the
preallocation_test
branch
from
74e9d59 to
a441a72
Compare
|
So it looks like currently we only need to wait a bit before destroy, otherwise I get the same BSOD as in #282 |
EchterAgo
force-pushed
the
preallocation_test
branch
4 times, most recently
from
b40bb5a to
b8e7f1c
Compare
I used this test when diagnosing openzfs#281 and think it would be a good addition. This tests file sizes when pre-allocating a file. Expected values have been confirmed on NTFS, ReFS and FAT32. This test currently does not work on a SMB share with the Samba server because it reports a fixed cluster size based on a configuration option instead of getting the correct value from the underlying file system. ksmbd does not have the same issue, it correctly gets the values from the file system. So far this is untested with Windows SMB shares. This also makes `tests.py` use some helpers from `utils.py`. Signed-off-by: Axel Gembe <[email protected]>
EchterAgo
force-pushed
the
preallocation_test
branch
from
b8e7f1c to
be98e09
Compare
|
So this should be ready, the failing tests are Ubuntu 22.04 which seems unrelated and @andrewc12 Did the |
I don't know if it ever did, I honestly cant remember |
andrewc12
merged commit 11cb959
into
openzfsonwindows:zfs-Windows-2.2.0-release
|
Yeah this is all good stuff. |
|
I was also planning to update the rest of the tests, but let me know if you need any help/additions or have any suggestions for things in I want to try to create tests for previous issues I've reported here and have since been fixed. |
I used this test when diagnosing #281 and think it would be a good addition. This tests file sizes when pre-allocating a file. Expected values have been confirmed on NTFS, ReFS and FAT32. This test currently does not work on a SMB share with the Samba server because it reports a fixed cluster size based on a configuration option instead of getting the correct value from the underlying file system. ksmbd does not have the same issue, it correctly gets the values from the file system. So far this is untested with Windows SMB shares. This also makes `tests.py` use some helpers from `utils.py`. Signed-off-by: Axel Gembe <[email protected]>
I used this test when diagnosing #281 and think it would be a good addition. This tests file sizes when pre-allocating a file. Expected values have been confirmed on NTFS, ReFS and FAT32. This test currently does not work on a SMB share with the Samba server because it reports a fixed cluster size based on a configuration option instead of getting the correct value from the underlying file system. ksmbd does not have the same issue, it correctly gets the values from the file system. So far this is untested with Windows SMB shares. This also makes `tests.py` use some helpers from `utils.py`. Signed-off-by: Axel Gembe <[email protected]>
I used this test when diagnosing #281 and think it would be a good addition.
This tests file sizes when pre-allocating a file.
Expected values have been confirmed on NTFS, ReFS and FAT32. This test currently does not work on a SMB share with the Samba server because it reports a fixed cluster size based on a configuration option instead of getting the correct value from the underlying file system. ksmbd does not have the same issue, it correctly gets the values from the file system. So far this is untested with Windows SMB shares.
This also makes
tests.pyuse some helpers fromutils.py.Note that currently the test does not pass yet because #281 needs more fixing.