| opendevreview | Harsh Ailani proposed openstack/cinder master: [IBM SVf Driver] Added support of clone operation for volumegroup https://review.opendev.org/c/openstack/cinder/+/938771 | 07:00 |
|---|---|---|
| opendevreview | Harsh Ailani proposed openstack/cinder master: [IBM SVf Driver] Added support for volumegroup-snapshot for temporary volumegroup https://review.opendev.org/c/openstack/cinder/+/926844 | 07:30 |
| *** jhorstmann is now known as Guest5975 | 07:51 | |
| *** geguileo is now known as Guest5985 | 09:58 | |
| *** jhorstmann is now known as Guest5997 | 13:46 | |
| sp-bmilanov | VolodymyrBoiko[m]: welcome! just in case you missed it, there is a "new driver checklist" in the docs (https://docs.openstack.org/cinder/latest/contributor/drivers.html#new-driver-review-checklist) you will also need a CI that can run integration tests with OpenStack that is connected to your storage system. I can share my experiences if you haven't yet gotten to setting a CI up | 13:50 |
| sp-bmilanov | it might seem there are a lot of things to consider upfront, but it is manageable | 13:51 |
| jbernard | jungleboyj rosmaita smcginnis tosky whoami-rajat m5z e0ne geguileo eharney jbernard hemna fabiooliveira yuval tobias-urdin adiare happystacker dosaboy hillpd msaravan sp-bmilanov Luzi sfernand: meeting in #openstack-meeting-alt \o/ | 14:00 |
| opendevreview | Luigi Toscano proposed openstack/cinder stable/2024.2: NFS update volume attachment format during volume snapshot https://review.opendev.org/c/openstack/cinder/+/939349 | 14:14 |
| opendevreview | Luigi Toscano proposed openstack/cinder stable/2023.2: NFS update volume attachment format during volume snapshot https://review.opendev.org/c/openstack/cinder/+/939350 | 14:14 |
| opendevreview | Luigi Toscano proposed openstack/cinder stable/2024.1: NFS update volume attachment format during volume snapshot https://review.opendev.org/c/openstack/cinder/+/939351 | 14:15 |
| rosmaita | tosky: do we need these backported so there can be tempest test coverage? | 14:24 |
| tosky | rosmaita: yes, but I guess we want to backport a relevant fix anyway, right? | 14:24 |
| rosmaita | tosky: good point | 14:25 |
| josephillips | hi | 15:04 |
| josephillips | simondodsley . for example netapp not support pergb QoS and they calculate the size of the volume and apply the QoS policy on the storage backend | 15:05 |
| simondodsley | josephilips netapp provide a key for this type of per gb qos because the NetApp platforms support this feature. Not all backends support it. In these case, eg caph, if you want per gb qos you have to revert to front-end | 15:05 |
| simondodsley | ok - that is interesting | 15:06 |
| josephillips | no ,. netapp don't support pergb on the backend side | 15:06 |
| simondodsley | what do they do when the volume size is changed as you can't dynamically change the qos | 15:06 |
| simondodsley | or can they do that as well? | 15:06 |
| josephillips | when a extend is executed on the resize you apply the new policy to ceph | 15:07 |
| josephillips | and netapp work on the same way | 15:07 |
| simondodsley | ok - i see | 15:07 |
| josephillips | my question is . i should use the same keys that is for frontend | 15:08 |
| whoami-rajat | so we will add manual code to the RBD driver to change iops limit when extend happens | 15:08 |
| josephillips | or i should use a different key like maxIOPSperGB | 15:08 |
| josephillips | when extend , create a new volume , retype etc. | 15:08 |
| josephillips | if the qos spec on backend total_iops_per_gb is set | 15:09 |
| whoami-rajat | ok, i think it's best to use same names for consistency, does netapp use same names or netapp specific ones? | 15:09 |
| josephillips | no | 15:09 |
| josephillips | netapp use maxIOPSperGB | 15:09 |
| simondodsley | is that because they can support lower limits as well, so there is an equivalent min? If not, then I would assume the qos limit is logically max anyway | 15:11 |
| josephillips | yep actually ceph can do that , but i can implement the _min too | 15:12 |
| josephillips | https://github.com/openstack/manila/blob/f89994e5d94ae96f7ab36842ee794d85b20e52b4/manila/share/drivers/netapp/dataontap/cluster_mode/lib_base.py#L1572C55-L1572C65 @simondodsley | 15:13 |
| josephillips | sprry wrong | 15:14 |
| josephillips | project | 15:14 |
| josephillips | https://github.com/openstack/cinder/blob/962fe29e778c58a8e90c78602e7249cb4b06e450/cinder/volume/drivers/netapp/utils.py#L319 @simondodsley | 15:14 |
| simondodsley | i see | 15:15 |
| josephillips | so what you guys think follow maxIOPSperGiB keys or keep exactly like the documentation of capacitive QoS but in this case for the backend? i mean i have to update the documentation too | 15:16 |
| whoami-rajat | since rbd/ceph is a community driver, it's best to to decide this in presence of team, maybe the upcoming midcycle in 2 weeks is a good place to have it where we can discuss this in detail | 15:18 |
| josephillips | alright what i will do is finish the modification because i already have implementet with total_iops_per_gb , in case is just only replace the keys | 15:19 |
| whoami-rajat | sounds good, thanks for working on this | 15:20 |
| josephillips | np | 15:25 |
| opendevreview | Volodymyr Boiko proposed openstack/cinder master: add volume driver for VASTData storage https://review.opendev.org/c/openstack/cinder/+/939005 | 15:35 |
| -opendevstatus- NOTICE: The paste service at paste.opendev.org will have a short (15-20) minute outage momentarily to replace the underlying server. | 17:07 | |
| opendevreview | Liron Kuchlani proposed openstack/devstack-plugin-nfs master: [DNM] test jobs https://review.opendev.org/c/openstack/devstack-plugin-nfs/+/939395 | 20:40 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!