- Group profile for transactional DB workload added
- Fixed the option names used in other group profiles
- Fixed the validation issues related to setting the options
glusterfs PR to support enable/disable of xlators
https://review.gluster.org/21813
Fixes: #1250
Signed-off-by: Aravinda VK <avishwan@redhat.com>
Moved lvm related functions from `$SRC/glusterd2/snapshot/lvm`
and `$SRC/plugins/device/deviceutils/` to `$SRC/pkg/lvmutils`
Also moved fs related functions from `plugins/deviceutils` to
`$SRC/pkg/fsutils`.
Fixes: #1187
Signed-off-by: Aravinda VK <avishwan@redhat.com>
The unit test (TestWriteStatedump) uses a filename derived from current
timestamp and the actual code (WriteStatedump) being tested also uses
filename of the same pattern but that one may have another timestamp as
that code executes slightly later. Hence, it can so happen that the file
name written by WriteStatedump() is different than the filename expected
by the test. As these two timestamps are at seconds level granularity,
we didn't see it fail earlier.
This change modifies the test to look for a pattern that resembles
statedump file.
Signed-off-by: Prashanth Pai <ppai@redhat.com>
A query string hash (QSH) is an important technique
to prevent URL tampering and secure HTTP requests,
When a Client makes a request without a QSH claim,
or with a bad QSH claim (one that doesn’t match
the request), then the glusterd2 will respond with
HTTP status 401 Unauthorized
Signed-off-by: Madhu Rajanna <mrajanna@redhat.com>
The err string returned by `exec.Command.{Run,Output}` will
not contains `stderr`, Output in log/CLI will just display `exited with
status 1`. This patch adds the content of `stderr` along with the error
message.
Signed-off-by: Aravinda VK <avishwan@redhat.com>
peer add in ETCD.
if peer address already present in store
update configuration on startup
this allows the users to use peer address
provided during peer add, for any other
operations(ex:- volume create,expand)
Signed-off-by: Madhu Rajanna <mrajanna@redhat.com>
If a glusterd2 instance is already processing a cluster membership
request (leave or join), return 409 (Conflict) to the other such
client requests immediately.
Signed-off-by: Prashanth Pai <ppai@redhat.com>
* Separated brick validation step from the step which modified it by
setting xattrs.
* Added rollback functions.
Signed-off-by: Prashanth Pai <ppai@redhat.com>
