| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|/ /
| |
| |
| |
| | |
Change-Id: Ic363aed46b73326bdbbffd85b077a13c92501f55
Signed-off-by: Michael Adam <madam@redhat.com>
|
| |
| |
| |
| |
| | |
Change-Id: I84b0399bad021fed29044c1946fdf27cccc2b6c8
Signed-off-by: Apeksha D Khakharia <akhakhar@redhat.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Improvements:
- Stop using 'create_mongodb_pod' function, it is replaced with
lightweight analog, which works ok on OCP3.11
- Pick up Glusterfs POD for recreation intelligently
in 'gluster POD failure' test case.
- Create secrets and storage classes from CLI, not using files.
Change-Id: I28d819574533573e5f397e2d2db6f0433a59eb99
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Improvements:
- Stop using 'create_mongodb_pod' function, it is replaced with
lightweight analog, which works ok on OCP3.11
- Pick up Glusterfs POD for recreation intelligently
in 'gluster POD failure' test case.
- Create secrets and storage classes from CLI, not using files.
Change-Id: I26ba5046014b84caacbf9ab94a2509b4324e112f
|
|\ \ |
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Blocked by BZ-1629889.
When referenced bug is fixed completely we should refactor this
test first, making it not dependent on the specific volume sizes
as it is now.
Change-Id: I1d4c4a0605fd800fa8450f92dab7b50e29856a5e
|
| |
| |
| |
| |
| |
| | |
testcases"
This reverts commit 0b6fcb9989287cc2748312dc59d52d38f7fda1ad.
|
| |
| |
| |
| | |
This reverts commit 014521179de579dc4744ab36efc0c2f8881e2f3c.
|
| |
| |
| |
| | |
Change-Id: I0d2de16928de8a533745876094c673152012d812
|
|/
|
|
|
| |
Change-Id: I083b5b60917ec18c77357d615f33e963ee58efea
Signed-off-by: vamahaja <vamahaja@redhat.com>
|
|
|
|
|
|
| |
Where we create POD with mounted arbiter volume and write data to it.
Change-Id: I3a757f7fd71a0fd3f9f7d3cea463bbba5e399ec4
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
List of enhancements:
- Add handler for SIGTERM, which allows us to stop POD execution
gracefully instead of timing out for grace period and killing it.
- Use separate data file name for each of the PODs based on their
unique names. It is much more useful approach when we create more
than 1 POD in a DC. Without it we write to one single file with locks,
with this change we write data to different files by each POD in
parallel.
Change-Id: If4bba1c017df1cf3240896522720fa08d5cf138b
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| | |
Making its 'test_volume_creation_no_free_devices' test
be aware of the actual available free space, which will make it work
much more reliably.
Change-Id: I2555ffeaad825b16ffee7ded067eab92d40519e4
|
|\ \
| |/
|/| |
|
| |
| |
| |
| |
| | |
Change-Id: I49932f43ac4e41f06c559d1d18d142b8a084278e
Signed-off-by: vamahaja <vamahaja@redhat.com>
|
|\ \ |
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Skip 2 tests in 'test_heketi_volume_operations' test module due to
existing unfixed bug: BZ-1629889
After fix of this bug, these 2 tests should be refactored the way,
where we leave Heketi cluster as was, not changing set of disks, as it
is now.
Change-Id: I020e7b91c41c3bd93cf1984f733adcdcbd361f92
|
|/
|
|
|
|
|
|
| |
The single test in this module passes only in small amount of Heketi
cluster states. So, refactor it completely to be reliable making it fail
only on proper errors.
Change-Id: I29251d4265ec5916d41bd3d69e21492f89097337
|
|
|
|
| |
Change-Id: I257f3e6a7389c9f15a4388c241e787782c186d14
|
|
|
|
|
|
|
| |
CNS-708,709,713,714,715
Change-Id: I61965e6d8b17b4bb49b10ebe26140945da9c4576
Signed-off-by: Apeksha D Khakharia <akhakhar@redhat.com>
|
|\ |
|
| |
| |
| |
| |
| | |
Change-Id: Ie3169e77e542201167fa57bee710e600b25b5a74
Signed-off-by: Apeksha D Khakharia <akhakhar@redhat.com>
|
|\ \ |
|
| |/
| |
| |
| |
| |
| |
| |
| | |
With additional logic for PVC deletion.
Make it also wait for deletion of PVC's undelying PV.
Doing so, we wait for deletion of volume from storage backend,
Change-Id: I2aab7ebb7924968a2d1825056d5747df5b816cd7
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Removing not used imports and aligning code to satisfy 79 symbols
per line.
all the pep8 errors and warnings can be seen using following command:
$ tox -e pep8
Change-Id: I409ce3fb547ed8891063f59d9218d8507e7bf6fc
|
| |/
|/|
| |
| |
| |
| |
| | |
Which allow to set and remove tags on Heketi node and devices.
Change-Id: Iaca652b059e58951f20bc975bf5724f6a5d0ca0e
Signed-off-by: nitin goyal <nigoyal@redhat.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Constant I\O will run on its PODs.
PODs are based on lightweight image called "cirros", which weighs about
10Mb. I\O will run on the mounted PV from PVC we specified in function
arguments.
Reuse this new function in couple of tests to show that it works.
Change-Id: I7ae6c252a35f566ec674c8f8c17847fc37b554b8
|
| |
| |
| |
| |
| |
| |
| |
| | |
In OCP3.11 we get additional field for "oc get pods -o wide" command,
so, update regex pattern to support new approach keeping compatibility
with OCP3.10- versions.
Change-Id: Ib16cdc1f584702cf829d5fc50e6dd87ef290e089
|
| |
| |
| |
| |
| | |
Change-Id: I52cb4e6360676db1b75764ec314ec53c017ae8d2
Signed-off-by: Apeksha D Khakharia <akhakhar@redhat.com>
|
|/
|
|
|
| |
Change-Id: I97eb3ae1d0af48d405a0187fe585e2732658f809
Signed-off-by: Apeksha D Khakharia <akhakhar@redhat.com>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Do following improvements to this function:
- Remove support of python Heketi client, because it is dead code.
It is dead code, because all of the tests which use this function,
explicitly depend on the CLI output, not being compatible with it.
- Remove useless 'json' arg, as CLI client does not support it for 'node
list' command.
- Parse list of node IDs in it, not delegating this to the consumer.
- Raise exception in case we failed to get list of nodes
And, of course, update all the places where this function is used.
Making those places use less code compared to previous variant.
Change-Id: I3f1ffe6082541e63dc3d638c19c54086aa12580f
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Also, 3 new library functions for:
- secret creation
- storage class creation
- pvc creation
Which use "stdin" creation approach without files.
Change-Id: Ifba768deba11048ede207d72af4d480898e8f5fc
|
|\ \
| |/
|/| |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
One single run of this func, depending on the size of iptables config
file on remote machine, could take minute or more for execution.
It is completely unacceptable just for updating one single file.
So, replace inline change of remote file with simple single operation of
write to it with whole set of needed inner data.
Having cluster of 2 nodes and trying to run tests which have this func in
its 'setUpClass' method, we could wait up to the 5 minutes just to reach
execution of first test. With growth of amount of cluster nodes this
waiting time would only lineary increase.
Change-Id: Ief0b4d3b879ddc9f65557df54beddbb51bc28ba5
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
In those tests, we scale down amount of heketi pods to 0 and then
scale it up back to amount of 1. And right after it, we perform request
for list of Heketi PODs, not waiting for Heketi POD appearing in list.
So, fix it by adding waiting for 2 sec which is enough time for Heketi
POD to appear in the POD list.
Change-Id: I5361d5c6eb98ac46b5f49a70edd0a88d550ab0b8
|
| |
| |
| |
| |
| | |
Change-Id: I78cb7989876e2f17cf1d9653723a160074050413
Signed-off-by: Apeksha D Khakharia <akhakhar@redhat.com>
|
|/
|
|
|
| |
Change-Id: Ic4ebb29be6070e982c53cfa63730e6dc4d955711
Signed-off-by: Apeksha D Khakharia <akhakhar@redhat.com>
|
|\ |
|
| |
| |
| |
| |
| |
| | |
Change-Id: Iedd408844beadd1a449fdb953640289044c1bfa7
Signed-off-by: Saravanakumar Arumugam <sarumuga@redhat.com>
Signed-off-by: Apeksha D Khakharia <akhakhar@redhat.com>
|
|\ \ |
|
| |/
| |
| |
| |
| |
| |
| |
| |
| | |
Also removing the skip decorator, the testcase
logic have no issues and it passes when executed
individually
Change-Id: Iede3f6b67ddd79d0efc673b2b3440fba392540d4
Signed-off-by: Apeksha D Khakharia <akhakhar@redhat.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Fix 2 autotests, remove skip decorator for them.
And rework logic of mongodb template uploading for mongodb pod,
so it takes couple of seconds instead of minutes.
Change-Id: Ib2b09364ae435b9784b76c2f2581c197128f9649
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: I84968bcfe19aa8af73f185be178079a8373915b3
Signed-off-by: Apeksha D Khakharia <akhakhar@redhat.com>
|
|\ \ \
| |_|/
|/| | |
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
To avoid errors trying to install latest dependencies using
unexpected pip version.
It allows us to run tests on RHEL7.4
Also, update 'README' file with actual information about running
commands in venvs.
Change-Id: I154be92cbaa46b7169d2a259acdb05d50448d570
|
|/
|
|
|
|
|
|
|
|
| |
With this change we do not face 'No such file or directory' error
anymore running tests using 'cnslibs' installed in system.
Without it we can run tests not facing such error only using 'tox'.
To have it applied we need to reinstall the 'cnslibs' package.
Change-Id: Ibf16ef6354210c4356e0c25f16582d0e73e0df13
|