It has been a process that has been evolving. We originally started with the
mysql, then dropped this as it did not satisfy cases when a different DB was
used. We then used memcached which was a disaster. The CPU of process/threads
that would take a lock would spike and this killed the system pe
Please ignore my second mail. I found the issue with the cleanup. Before
issuing "multipath -f ", os-brick executes "multipath -r" (reload). This
command is triggering an error in multipathd, which erases
/dev/mapper/, and subsequent check for this folder in os-brick
fails.
The error in multipathd
Regarding the cleanup in detach, I found that the python-os-brick package
in Ubuntu 16 is rather old: 1.2.0
I found that the function find_multipath_device
in os_brick/initiator/linuxscsi.py was refactored her
https://github.com/openstack/os-brick/commit/3ea86f7d605d7109f4e13c1566ba1a14dc856443
I
Hi,
Is someone using iscsi multipath in Ubuntu 16 / Mitaka? Things that work OK
for us in Ubuntu 14 / Icehouse seem to be broken.
Basically, the WWID of the volume is not automatically placed in
/etc/multipath/wwids, and if we add it manually, we can attach with
multipath, but the detach never cl
In case of any heat error due to either bad template or error from the
nova/neutron, the heat client returns a description indicating the cause of
the error.
However, if ReST API is used, only 500 Internal Server Error response is
returned and the consumer has no clue of the cause of the error.
Th
Out of curiosity which backend is the vmware_nsx code defaulting to (or
recommending?)?
-Josh
Gary Kotton wrote:
Hi,
In the vmware_nsx code base we make use of tooz extensively and the
results have been very good. Just depends on what type of backend you use J
We wrote a wrapper that enables
Hi,
In the vmware_nsx code base we make use of tooz extensively and the results
have been very good. Just depends on what type of backend you use ☺
We wrote a wrapper that enables the deployer to decide on what type of locking
to use. This can be seen at:
https://github.com/openstack/vmware-nsx/
Hi All,
Please let me know if any examples are there with Tooz library for
distributed locking.
Just wanted to know even though we have database locks to prevent
concurrent updates and if applications are written well enough to handle
STALE updates, does DLM is needed for openstack services like
