| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
Dict attribute called "iteritems()" is not supported in the py3.
So, replace it's usage with another similar attr called "items()".
Change-Id: I130b7f67f0a2d5da5ed6c3d792f5ff024ba148f4
Signed-off-by: Valerii Ponomarov <kiparis.kh@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The tearDown function was incomplete where, since the test case has
enabled the shared storage, it needs to be disabled as part of the
tear down.
This fix includes the addition of the necessary tearDown parts along
with some cosmetic changes.
Change-Id: Id421c840a1c7606ecf185c9520ca436d47911f45
Signed-off-by: Vinayak Papnoi <vpapnoi@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As the below mentioned bug is fixed resubmitting testcase:
https://bugzilla.redhat.com/show_bug.cgi?id=1690753
Test case:
-> Creating two volumes and starting them, stop the second volume
-> Set the server quorum and set the ratio to 90
-> Stop the glusterd in one of the node, so the quorum won't meet
-> Peer probing a new node should fail
-> Volume stop will fail
-> Volume delete will fail
-> Volume reset will fail
-> Start the glusterd on the node where it is stopped
-> Volume stop, start, delete will succeed once quorum is met
Change-Id: Ic9dea44364d4cb84b6170eb1f1cfeff1398b7a9b
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
| |
Change-Id: I12b5586bdcef128df64fcd8a0ba80f193395f313
Co-authored-by: Vijay Avuthu <vavuthu@redhat.com>
Signed-off-by: Vijay Avuthu <vavuthu@redhat.com>
Signed-off-by: Milind Waykole <milindwaykole96@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Test case:
* Enable quota on the volume.
* Set hard and soft time out to zero.
* Create some files and directories from mount point
so that the limits are reached.
* Perform add-brick operation on the volume.
* Start rebalance on the volume.
* While rebalance is running, kill one of the bricks of
the volume and start after a while.
* While rebalance + self heal is in progress,
create some more files and directories from the mount point until
limit is hit.
Change-Id: Ic7d2ac92b4e132ab1018242c17bed5e888e86cf3
Co-authored-by: Sanoj Unnikrishnan <sunnikri@redhat.com>
Signed-off-by: Sanoj Unnikrishnan <sunnikri@redhat.com>
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Test statvfs calls return appropriate avaialable size with quota.
* Enable Quota
* Save the result from statvfs call
* Set Quota limit of 1 GB on the root of the volume
* Validate statvfs call honors quota
* Remove quota limit from the Volume
* Validate statvfs call reports old value of avialable space
Change-Id: I5f6271e0acdba13d483eb321f62ca9fdc5360859
Signed-off-by: Sanoj Unnikrishnan <sunnikri@redhat.com>
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
| |
Adding test case to check posix compliance
for nfs ganesha
Change-Id: I8188fba9245cfe27ae30b7818cfe7d2b624e9d87
Signed-off-by: Ambarish Soman <asoman@redhat.com>
Signed-off-by: Arjun Sharma <arjsharm@redhat.com>
|
|
|
|
| |
Change-Id: Iec8471de0add8cb6eaf6c80fb24c631e992aad4d
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Earlier in the testcase we were turning off shd
which is not correct and we have to turn off only
client side heal options as mentioned below
metadata-self-heal
entry-self-heal
data-self-heal
After renaming files we have to turn on these options
while doing a look up from client
Change-Id: I8c76abb8e79620c412e5991f5d8255b6b2a850e8
Signed-off-by: Milind Waykole <milindwaykole96@gmail.com>
|
|
|
|
|
| |
Change-Id: I2afd326fe0df1959a8fd20bc7325159688c1a73e
Signed-off-by: Jilju Joy <jijoy@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Add-brick command failure is leaving the directories created
on the backend bricks which is resulting in the failure of the
subsequent cases.
Added some changes to clean the bricks.
Change-Id: I108efbcaef2010f6fd52c334446059f96fff3741
Signed-off-by: Prasad Desala <tdesala@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
-> Create volume
-> Stop the volume
-> Enabling serve quorum
-> start the volume
-> Set server quorum ratio to 95%
-> Stop the glusterd of any one of the node
-> Perform rebalance operation operation
-> Check gluster volume status
-> start glusterd
Change-Id: I3bb42a83414dbcabdc61178e11d584eaf90c3b40
Signed-off-by: Rajesh Madaka <rmadaka@redhat.com>
|
|
|
|
|
|
|
|
|
| |
In a few testcases in glusterd count++ is missing due to which the
testcase in an infinite loop. Fixing that and sending patch.
Change-Id: I56a355f6ea3ae79231e09d7aee80031da3ebec52
Signed-off-by: yinkui <13965432176@163.com>
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
| |
Change-Id: I7f8769defd34d55d8eec720c40ed55e69523f917
Signed-off-by: Anees Patel <anepatel@redhat.com>
Signed-off-by: milindw96 <milindwaykole96@gmail.com>
|
|
|
|
|
|
|
|
|
| |
Since the NfsGaneshaVolumeBaseClass has been removed
the ACL test needs to be modified to use the alternative
method to the above mentioned class.
Change-Id: I398dfef6dd334a3e3f3871d44705af312d81318a
Signed-off-by: Arjun Sharma <arjsharm@redhat.com>
|
|
|
|
|
|
|
|
|
| |
* The issue with disperse and distributed-disperse volume types are now fixed.
* Reference : Bugzilla 1663375
* Using mount object instead of clients list for setting authentication
Change-Id: I914cee7fb790dc65e947e0b6db40d02e23575e65
Signed-off-by: Jilju Joy <jijoy@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Earlier in nfs_ganesha_ops, two library functions
enable_acl, disable_acl were implemented and later
with changes to NFS Ganesha Base class both the
functions are removed and made into single function
set_acl.
Change-Id: I5456adeeffa49c35a5ea19c8d11272f91ec4bdbf
Signed-off-by: Bala Konda Reddy M <bmekala@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Since egrep treats a pattern as extended regular expression
it is more suited to be used in this case. The current usage
of grep fails to identify the expression after the first pipe.
Change-Id: Id02f884356048c56066a406937f8d86456ee8bd7
Signed-off-by: Arjun Sharma <arjsharm@redhat.com>
|
|
|
|
|
|
|
| |
Change-Id: I5c9c6ce48a973c6c902186f4e62afda196faf442
Signed-off-by: vivek das <vdas@redhat.com>
Signed-off-by: srivickynesh <sselvan@redhat.com>
Signed-off-by: vivek das <vdas@redhat.com>
|
|
|
|
|
|
|
|
| |
NfsGaneshaVolumeBaseClass and NfsGaneshaIOBaseClass are
essentially not needed and act as redundant.
Change-Id: I45f95e018daa4bf9575f4e831111f91615085bdc
Signed-off-by: Arjun Sharma <arjsharm@redhat.com>
|
|
|
|
|
|
|
| |
The method 'refresh_config' can be used instead of 'run_refresh_config'
Change-Id: I6c70b4dff1fd4531fdf6eaa4f044d2d1992d8151
Signed-off-by: Jilju Joy <jijoy@redhat.com>
|
|
|
|
|
|
| |
Change-Id: Ifef2ffe022accf59edcbc949c505f47931b19fe4
Signed-off-by: Anees Patel <anepatel@redhat.com>
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
| |
Change-Id: I5372caf852b41e127c46f6afa697435dcde9dbf6
Signed-off-by: Jilju Joy <jijoy@redhat.com>
|
|
|
|
|
| |
Change-Id: I40e36e1d4d404fcb9709fa9e50d33cad77172350
Signed-off-by: Jilju Joy <jijoy@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Steps
1. Start IO on mount points
2. Add bricks to expand the volume
3. Start rebalance and wait for its completion
4. Validate IOs
Change-Id: I97c09c9fc226afeff1446d225959730715f89aef
Signed-off-by: Jilju Joy <jijoy@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Steps
1. Start IO on mount points
2. Create another volume 'volume_new'
3. Export volume_new through nfs-ganesha
4. Mount the volume on clients
Change-Id: I2c4fe59e9a85e6668672d31a0a6c27d11c7f03f8
Signed-off-by: Jilju Joy <jijoy@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
After modifying file_dir_ops.py as per patch [1]
we would need to install numpy on clients. Hence
making the changes to the Readme.md.
[1]https://review.gluster.org/#/c/glusto-tests/+/23143/
Change-Id: Id3d6f4972e24317e470a8c68ed7172c516f53fc8
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
At present there are 2 different functions for prerequisites
and create session merging them and adding checks to make the
functions more effective when called from a conftest.py file.
By doing so we also avoid the possibility of running the same
function twice which might lead to problems.
For example when commands like:
$ gluster system:: execute gsec_create
it would make 2 entries in authorized keys files leading
to workers getting into faulty state.
Note:
This patch should only be merged after
merging the below mentioned patches:
[1] https://review.gluster.org/#/c/glusto-tests/+/23291/
[2] https://review.gluster.org/#/c/glusto-tests/+/23293/
Change-Id: I3f8cda7b589c86de4dab8c66d033b364dd495b9b
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
THIS PATCH IS DEPENDENT ON THE BELOW PATCH:
https://review.gluster.org/#/c/glusto-tests/+/23291/
Removing the below functions as a more generic function
is added in lib_utils:
georep_ssh_keygen
georep_ssh_copyid
georep_groupadd
georep_geoaccount
georep_geoaccount_setpasswd
Adding georep_mountbroker_remove_user to
remove users from mountbroker which
will be needed in teardown.
Change-Id: I5b484b93cbc1660cdf223cfa4bac2a32acc8fc77
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adding the following functions tot lib_utils.py:
1.ssh_keygen(): Creates a pair of ssh private
and public key if not present.
2.set_passwd(): Sets password for a given username.
3.ssh_copy_id(): Copies the default ssh public key.
4.group_add(): Creates a group in all the servers.
Modifying add_user() to create and add user to a
specific group. Also modifying the code in add_user()
to take list of servers.
Change-Id: I58cdef967293204073aa912fa98adb9606f6885d
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
| |
Change-Id: I3fb826bd0ecbe46bee4b9f8594b23f16921adbec
Signed-off-by: Arjun Sharma <arjsharm@redhat.com>
|
|
|
|
|
|
|
|
|
| |
With redundancy count as negative or disperse count
with negative(different permutations) and disperse
data count equal to disperse count
Change-Id: I761851c64833256532464f56a9a78e20ceb8a4e1
Signed-off-by: ubansal <ubansal@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The test case 'test_client_side_quorum_with_fixed_validate_max_bricks'
does not have a tearDown part where the volume options which have been
set inside the test case have not been reset to default.
The library function 'set_volume_options' was being imported from a
wrong library. This fix includes this change along with the tearDown
steps.
Change-Id: Ic57494e7a7e8a25303b7979f98cc2dfbc9a7d7b6
Signed-off-by: Vinayak Papnoi <vpapnoi@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For each node should have one self heal daemon and
one quota daemon should be up running
that means total number of self heal daemon and quota daemon
to be up and ruuning is (number of nodes *2),
in code i am checking that count should
be equalent to (number of nodes * 2)
Change-Id: I79d40467edc255a479a369f19a6fd1fec9111f53
Signed-off-by: Rajesh Madaka <rmadaka@redhat.com>
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The ExecutionError msg at line 66 should be "failed to clean-up
volume" because it is under tearDownClass. The g.log.info msg
at line 123 should be "Checking if gfid xattr of directories is
displayed and is same on all the bricks on the server node" as the
code below it checks for the gfid xattr on the bricks on server
node and not if it is displayed on mount point.
Change-Id: If4e20e3487a44c1cc7047504d19cc9859424ccd4
Signed-off-by: sayaleeraut <saraut@redhat.com>
|
|
|
|
|
| |
Change-Id: I240ecb0b4a9c99134b7a5cd237a59c2857d0fb7b
Signed-off-by: yinkui <13965432176@163.com>
|
|
|
|
|
|
|
|
|
|
| |
According to generate_io.py the function
"def check_if_percent_to_fill_or_timeout_is_met" will
never return false, the "flag" empty or not empty also
return true.
Change-Id: I2d70b2987112bd3a6404e5f13f0c6d320d01e7fc
Signed-off-by: yinkui <13965432176@163.com>
|
|
|
|
|
|
|
| |
the log.
Change-Id: I361b2c59108b19480906f6dfd49b023ed1eb05cd
Signed-off-by: yinkui <13965432176@163.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
After changing the type of unit file from INFO
to DEBUG. Performing daemon reload. Earlier using
running commands continuosly to generated debug
messages instead of running continuosly, restarted
glusterd in one of the nodes so that while handshake
the logs will be in Debug mode. After validating
reverting back the unit file to INFO and daemon
reload
Change-Id: I8c99407eff2ea98a836f37fc2d89bb99f7eeccb7
Signed-off-by: Bala Konda Reddy M <bmekala@redhat.com>
|
|
|
|
|
|
| |
Change-Id: I4d056b94b4ea59beee7eb24e7e5d5f65d7256b4a
Signed-off-by: Vitalii Koriakov <vkoriako@redhat.com>
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
| |
Made the IO process more efficient for better performance,
as lots of randomizations and conditionals within loops made the whole
process very slow.
Change-Id: Ie99340d6e5fae7ee88198a646f5b46f868c816ae
Signed-off-by: hadarsharon <hsharon@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
What is crefi?
Crefi is a Python command-line tool to create multi
threaded workload on a filesystem and do file operations
on created data. It provides basic operations to create files
of different sizes, different types, in different directory
structures, along with creating symlinks, hardlinks to files,
and also rename, truncate, chmod, chown, chgrp and
setxattr on the created files. Typical use involves creating
large number of files of different sizes, over different layout,
and do different operations on the created files.
Which components need this tool?
- Geo-rep
- AFR
Changes needed in infrastructure:
https://github.com/gluster/centosci/commit/1fe330d34a62b56438f6eb86538286962b1abd90
Change-Id: I1e579f6f1f2c6ef0d018c055234abbf0f147e621
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Verifying the glusterfind functionality with renames of files.
* Create a session on the volume
* Create various files from mount point
* Perform glusterfind pre
* Perform glusterfind post
* Check the contents of outfile
* Rename the files created from mount point
* Perform glusterfind pre
* Perform glusterfind post
* Check the contents of outfile
Files renamed must be listed
Change-Id: Ib7682e86d59f0519b267ec01cda999920a30de86
Signed-off-by: Vinayak Papnoi <vpapnoi@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Verifying the glusterfind functionality with deletion of files.
* Create a session on the volume
* Create various files from mount point
* Perform glusterfind pre
* Perform glusterfind post
* Check the contents of outfile
* Delete the files created from mount point
* Perform glusterfind pre
* Perform glusterfind post
* Check the contents of outfile
Files deleted must be listed
Change-Id: I2ee05a2c97983fb521648e372db21c8361a2c835
Signed-off-by: Vinayak Papnoi <vpapnoi@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
There were a large number of mistakes due to which
if we try to run the above playbook it failed on
the first task itself. As well as the format of
the playbook is as per a older version of ansible.
Fixing and submitting a patch for the same.
Change-Id: Ia8bd3264c290ff7abd12a9fa325d40673791ea11
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Steps:
-> Enable a shared storage
-> Disable a shared storage
-> Create volume of any type with
name gluster_shared_storage
-> Disable the shared storage
-> Check, volume created in step-3 is
not deleted
-> Delete the volume
-> Enable the shared storage
-> Check volume with name gluster_shared_storage
is created
-> Disable the shared storage
Change-Id: I1fd29d51e32cadd7978771f4a37ac87176d90372
Signed-off-by: Gaurav Yadav <gyadav@redhat.com>
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
| |
The test case makes use of 'is_shared_volume_unmounted' function
which is redundant as the purpose can be solved via a very similar
function 'is_shared_volume_mounted'. This change has been implemented.
There are various cosmetic changes that have been implemented as well.
Change-Id: I560b464b4bcc436658db49c0a5ed8c7aadacfb6a
Signed-off-by: Vinayak Papnoi <vpapnoi@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The test case 'test_client_side_quorum_with_fixed_for_cross3' does not
include the tearDown part where the volume options which have been set
inside the test case have to be reset to default.
This fix includes the necessary tearDown steps along with a few
cosmetic changes.
Change-Id: I86187cef4523492ec97707ff93d0eca365293008
Signed-off-by: Vinayak Papnoi <vpapnoi@redhat.com>
|
|
|
|
|
|
| |
Change-Id: I0f32dc73d960f8bfeb662643e736291ffb74cb81
Signed-off-by: Roch-elle <rallan@redhat.com>
Signed-off-by: kshithijiyer <kshithij.ki@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The line 316 of heal_libs.py library contains a function called
'wait_for_self_heal_daemons_to_be_online' which makes use of the
'volname' and 'timeout' arguements for string formatting.
The method used for the string formatting was wrong which causes
an error 'TypeError: not enough arguments for format string'.
Due to this, all the test cases which make use of heal_libs.py
will result in a TypeError scenario.
This patch will fix the TypeError for the string formatting.
Change-Id: I73abe573d2ccb60b74ccd6ae268b950640d75600
Signed-off-by: Vinayak Papnoi <vpapnoi@redhat.com>
|