Friday, 2026-05-15

srinesh[m]carloss: gouthamr Sure, for now I will try to go with the approach where in case user tries to set access rules for mountable snapshot, we will error the operation and report that it is not supported through user messages.05:39
srinesh[m]In case a separate capability for snapshot allow/deny like the goutham mentioned were to be implemented, would you expect us to implement it. Or it will be taken up as part of this or next release.11:06
opendevreviewGoutham Pacha Ravi proposed openstack/manila-tempest-plugin master: Re-enable IPv6 tests  https://review.opendev.org/c/openstack/manila-tempest-plugin/+/89521114:04
hong-p[m]carloss: hi I prepare the spec for availability_zone filter. could you see this? https://review.opendev.org/c/openstack/manila-specs/+/98865015:06
opendevreviewJonathan Koerber proposed openstack/manila master: api: Add Schema to share: manage share POST v2/share/manage  https://review.opendev.org/c/openstack/manila/+/98878616:27
-opendevstatus- NOTICE: The Gerrit service on review.opendev.org will be offline momentarily while we restart it onto a new patch release18:04
carloss> In case a separate capability for snapshot allow/deny like the goutham mentioned were to be implemented, would you expect us to implement it. Or it will be taken up as part of this or next release.18:10
carlosssrinesh[m] I think this release would be better, considering that it would already help erroring out earlier18:10
carlosshong-p[m]: sure, will take a look18:10
gouthamrhong-p[m]: carloss: i think the AZ filter not working as intended is a bug18:15
carlossyeah, it should be already there18:16
carlossI could not find it in the API though18:21
carlosshowever, we do allow az filtering for services18:21
gouthamryeah looks like it's an omission18:21
gouthamri'd be inclined to treat this as a bug18:22
carlossI feel like it's a bug though18:22
carlossyeah18:22
carlossthere's no reason why not to allow it18:22
gouthamrtoday, sending the "availability_zone" as a filter will be ignored18:22
gouthamrso you'll get a list of shares regardless of "availability_zone", when we fix this, the users will get a filtered subset18:29
gouthamrwe've treated filter additions as API behavior changes warranting a microversion bump18:29
gouthamrthis has been contentious18:29
gouthamrother teams have a policy that they published stating that they consider things like this as bug fixes and will still backport without a microversion bump18:30
gouthamrfor this issue, i feel the approach other services took seems saner 18:30
carlossagreed18:34
carlossthe only downside with that is that if a non-existent az is provided at the time of the request, currently the filter is just dropped18:35
carlossand life goes on18:35
carlosswith this change, if it's an non-existent az that is part of the request, we might need to raise an exception18:36
carlosswhich changes the behavior of the API18:36
carlossunless we ignore the failure and life goes on as it does today18:36
gouthamrdo we?18:36
gouthamri mean.. today if you filter by a non-existent name, you would get an empty list?18:37
carlossi might be trippin', but confirming18:37
carlosstoday, if we filter by a non-existent name, or if we just set *any* azs, we'll get the whole shares list18:43
carlossand it won't filter18:43
carlosshttps://github.com/openstack/manila/blob/master/manila/api/v2/shares.py#L146818:43
carlossshare's api _get_share_search_options() method does not have availability zone listed there18:43
carloss> the only downside with that is that if a non-existent az is provided at the time of the request, currently the filter is just dropped and life goes on18:44
carlossmeaning this is correct indeed18:44
carlosshttps://paste.opendev.org/show/boQhPBYww46fpvzCcCJl/18:46
carlossmaybe we should just preserve the behavior if we're treating this as a bugfix18:46
carlosswe ack that an availability zone is part of the request and try to filter shares with it. if it's non-existent, we just return the list of shares if nothing happened18:46
carlosswe took a similar approach for service filters: https://github.com/openstack/manila/blob/master/manila/api/v2/services.py#L8518:48
carlossbut in this case instead, we return an empty list18:48
carlossI'd be okay following the same approach18:49
gouthamr+118:50
gouthamrand probably wrote this down in the contrib guide18:50
gouthamr"how to handle filters consistently"18:51
opendevreviewGoutham Pacha Ravi proposed openstack/manila stable/2025.2: Support IPv6 wildcard ::/0 in Ganesha  https://review.opendev.org/c/openstack/manila/+/98882020:02
opendevreviewGoutham Pacha Ravi proposed openstack/manila stable/2025.1: Support IPv6 wildcard ::/0 in Ganesha  https://review.opendev.org/c/openstack/manila/+/98882120:03
opendevreviewGoutham Pacha Ravi proposed openstack/manila-tempest-plugin master: [DNM] Test IPv6 wildcard ::/0 in CephFS NFS  https://review.opendev.org/c/openstack/manila-tempest-plugin/+/98882421:45
opendevreviewGoutham Pacha Ravi proposed openstack/manila-tempest-plugin master: [DNM] Prove ::/0 bug in CephFS NFS (no fix)  https://review.opendev.org/c/openstack/manila-tempest-plugin/+/98882521:48
opendevreviewGoutham Pacha Ravi proposed openstack/manila-tempest-plugin master: Re-enable IPv6 tests  https://review.opendev.org/c/openstack/manila-tempest-plugin/+/89521122:05
opendevreviewGoutham Pacha Ravi proposed openstack/manila-tempest-plugin master: [DNM] Test IPv6 wildcard ::/0 in CephFS NFS  https://review.opendev.org/c/openstack/manila-tempest-plugin/+/98882422:05
opendevreviewGoutham Pacha Ravi proposed openstack/manila-tempest-plugin master: Re-enable IPv6 tests  https://review.opendev.org/c/openstack/manila-tempest-plugin/+/89521122:05
opendevreviewGoutham Pacha Ravi proposed openstack/manila-tempest-plugin master: [DNM] Prove ::/0 bug in CephFS NFS (no fix)  https://review.opendev.org/c/openstack/manila-tempest-plugin/+/98882522:05

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!