February 2020 - openstack-discuss

[operators][cinder] RSD and Swordfish drivers being abandoned
by Brian Rosmaita 13 Feb '20

13 Feb '20

Greetings operators, The current maintainers of the Intel Rack Scale Design (RSD) driver announced at this week's Cinder meeting that the driver will be marked as UNSUPPORTED in Ussuri and that the driver will no longer be maintained. The Swordfish driver, which was under active development, but not yet merged to Cinder, is also being abandoned as well as the os-brick modifications to support NVMeoF. Finally, the third-party CI system for the RSD driver is being dismantled. We want you to be aware of these changes, but more optimistically, we are hoping that there would be some interest in the community in picking these up. If you are or know of such a person, please contact us right away so that we can get a technology transfer going before the current development group has been dispersed to other projects. For more information, please contact: Ivens Zambrano (ivens.zambrano(a)intel.com / IRC:IvensZambrano) David Shaughnessy (david.shaughnessy(a)intel.com / IRC: davidsha) thanks, brian

1 0

[qa][ptg] QA PTG, Vancouver Planning
by Ghanshyam Mann 12 Feb '20

12 Feb '20

Hello Everyone, I have started the Vancouver PTG planning for QA[1]. The very first step is to reserve the space based on number of attendees. For that, please write your name in the etherpad. I know it might not be confirmed yet or the travel process has not started, still, you can write your probability of attending. It is not necessary to be 100% in QA but if you are planning to spend some time in QA PTG, still write your name with comment. [1] https://etherpad.openstack.org/p/qa-victoria-ptg -gmann

1 0

[cinder][ptg] Victoria (Vancouver 2020) PTG headcount
by Brian Rosmaita 12 Feb '20

12 Feb '20

I need to get a rough headcount to submit to the PTG planning committee. Please take a minute to put your name on this etherpad if you think/hope/plan to attend (no commitment--we just want to increase the likelihood that we'll have enough seats for the Cinder team): https://etherpad.openstack.org/p/cinder-victoria-ptg-planning While you're on that etherpad, if you have a topic idea already, feel free to add it. I need the headcount info by 19:00 UTC on 28 February. (But do it now before you forget.) thanks, brian

1 0

[largescale-sig] Meeting summary and next actions
by Thierry Carrez 12 Feb '20

12 Feb '20

Hi everyone, The Large Scale SIG held a meeting today. We had a small attendance, and blame holidays for that. But you can catch up with the summary and logs of the meeting at: http://eavesdrop.openstack.org/meetings/large_scale_sig/2020/large_scale_si… We started with a presentation by oneswig of the integration of HAProxy telemetry via Monasca on a Kolla-Ansible-based deployment, and how that can help us derive latency information based on the API server called, the HTTP method and the HTTP code result across all HAProxy-fronted servers in a deployment. Slides are available at: http://www.stackhpc.com/resources/HAproxy-telemetry.pdf Regarding progress on our "Documenting large scale operations" goal, some doc links were proposed and all seem relevant to our collection. On the "Scaling within one cluster, and instrumentation of the bottlenecks" goal, the draft of the oslo.metrics spec got a first round of comments. masahito will refresh it based on those initial comments, but it would be good for SIG members to also weigh in at that early stage. In other news, we discussed the SIG's involvement in the "Large-scale Usage of Open Source Infrastructure Software" track at the OpenDev event in Vancouver in June. We plan to have SIG members directly involved and have discussions around our two goals for 2020. TODOs between now and next meeting: - masahito to update spec based on initial feedback - everyone to review and comment on https://review.opendev.org/#/c/704733/ The next meeting will happen on February 26, at 9:00 UTC on #openstack-meeting. Cheers, -- Thierry Carrez (ttx)

1 0

[scientific-sig] No Scientific SIG meeting today
by Stig Telfer 12 Feb '20

12 Feb '20

Hi All - Unfortunately there will not be a Scientific SIG IRC meeting today, owing to other commitments and availability. Apologies, Stig

1 0

How to Migrate OpenStack Instance from one Tenant to another
by Pradeep Antil 12 Feb '20

12 Feb '20

Hi Folks, I am using Mitaka OpenStack and VLAN as network type driver for tenant VMs. Initially i have provisioned all the VMs (Including Customer VMs) inside the admin. I want to segregate my internal VMs and Customer VMs, for this to accomplish i have created different tenant names and now i want customer VMs from admin tenant to Customer tenant, On Google , i have a found a script which migrates the VMs but that script doesn't update network and security group details, Can anyone help me and suggest what steps needs to executed to update network , security group and attached Volumes ..? Below is the exact script: #!/bin/bash for i do if [ "$i" != "$1" ]; then echo "moving instance id " $i " to project id" $1; mysql -uroot -s -N <<query use nova; update instances set project_id="$1" where uuid="$i"; query else #get project id of the instance before update old_proj_id=$(mysql -uroot -s -N <<query use nova; select project_id from instances where uuid="$2"; query ) #get user id of the instance before update user_id=$(mysql -uroot -s -N <<query use nova; select user_id from instances where uuid="$2"; query ) fi done echo "quota_resync " $old_proj_id; ./quota_resync.sh $old_proj_id $user_id; echo "quota_resync " $1; ./quota_resync.sh $1 $user_id; root@xxxx:/mnt# cat quota_resync.sh #!/bin/bash count=$(mysql -uroot -s -N <<query use nova; select count(*) from instances where project_id="$1" and vm_state="active"; query ) memory=$(mysql -uroot -s -N <<query use nova; select sum(memory_mb) from instances where project_id="$1" and vm_state="active"; query ) vcpu=$(mysql -uroot -s -N <<query use nova; select sum(vcpus) from instances where project_id="$1" and vm_state="active"; query ) echo "count" $count; echo "vcpus" $vcpu; echo "Ram" $memory; update=$(mysql -uroot -s -N <<query use nova; update quota_usages set in_use=$count where project_id="$1" and user_id="$2" and resource="instances"; update quota_usages set in_use=$memory where project_id="$1" and user_id="$2" and resource="ram"; update quota_usages set in_use=$vcpu where project_id="$1" and user_id="$2" and resource="cores"; query ) root@xxxx:/mnt# Script Execution Method : # /vm_move.sh "destination_tenant_id" "VM_id" -- Best Regards Pradeep Kumar

1 0

Re: [ospurge] looking for project owners / considering adoption
by Artem Goncharov 12 Feb '20

12 Feb '20

Hi, I am thinking to submit (not found possibility so far) a forum-like session for that again in Vancouver, where I would present current status of implementation and we can plan further steps. Unfortunately I have still no confirmation from my employer, that I will be allowed to go. Any ideas/objections? Regards, Artem > On 3. Nov 2019, at 02:34, Tobias Rydberg <tobias.rydberg(a)citynetwork.eu> wrote: > > Hi, > > Sounds really good Artem! Will you be at the session at the Summit? If not, I will bring the information from you to the session... > > Cheers, > Tobias > > Tobias Rydberg > Senior Developer > Twitter & IRC: tobberydberg > > www.citynetwork.eu <http://www.citynetwork.eu/> | www.citycloud.com <http://www.citycloud.com/> > > INNOVATION THROUGH OPEN IT INFRASTRUCTURE > ISO 9001, 14001, 27001, 27015 & 27018 CERTIFIED > On 2019-11-02 16:26, Artem Goncharov wrote: >> Hi Tobby, >> >> As I mentioned, if Monty does not start work, I will start it in few weeks latest (mid November). I need this now in my project, therefore I will be definitely able to spend time on implementation in both SDK and OSC. >> >> P.S. mailing this to you, since I will not be on the Summit. >> >> Regards, >> Artem >> >>> On 2. Nov 2019, at 09:19, Tobias Rydberg <tobias.rydberg(a)citynetwork.eu <mailto:tobias.rydberg@citynetwork.eu>> wrote: >>> >>> Hi, >>> >>> A Forum session is planned for this topic, Monday 11:40. Suites perfect to continue the discussions there as well. >>> >>> https://www.openstack.org/summit/shanghai-2019/summit-schedule/events/24407… <https://www.openstack.org/summit/shanghai-2019/summit-schedule/events/24407…> >>> BR, >>> Tobias >>> >>> Tobias Rydberg >>> Senior Developer >>> Twitter & IRC: tobberydberg >>> >>> www.citynetwork.eu <http://www.citynetwork.eu/> | www.citycloud.com <http://www.citycloud.com/> >>> >>> INNOVATION THROUGH OPEN IT INFRASTRUCTURE >>> ISO 9001, 14001, 27001, 27015 & 27018 CERTIFIED >>> On 2019-10-30 15:43, Artem Goncharov wrote: >>>> Hi Adam, >>>> >>>> Since I need this now as well I will start working on implementation how it was agreed (in SDK and in OSC) during last summit by mid of November. There is no need for discussing this further, it just need to be implemented. Sad that we got no progress in half a year. >>>> >>>> Regards, >>>> Artem (gtema). >>>> >>>>> On 30. Oct 2019, at 14:26, Adam Harwell <flux.adam(a)gmail.com <mailto:flux.adam@gmail.com>> wrote: >>>>> >>>>> That's too bad that you won't be at the summit, but I think there may still be some discussion planned about this topic. >>>>> >>>>> Yeah, I understand completely about priorities and such internally. Same for me... It just happens that this IS priority work for us right now. :) >>>>> >>>>> >>>>> On Tue, Oct 29, 2019, 07:48 Adrian Turjak <adriant(a)catalyst.net.nz <mailto:adriant@catalyst.net.nz>> wrote: >>>>> My apologies I missed this email. >>>>> >>>>> Sadly I won't be at the summit this time around. There may be some public cloud focused discussions, and some of those often have this topic come up. Also if Monty from the SDK team is around, I'd suggest finding him and having a chat. >>>>> >>>>> I'll help if I can but we are swamped with internal work and I can't dedicate much time to do upstream work that isn't urgent. :( >>>>> >>>>> On 17/10/19 8:48 am, Adam Harwell wrote: >>>>>> That's interesting -- we have already started working to add features and improve ospurge, and it seems like a plenty useful tool for our needs, but I think I agree that it would be nice to have that functionality built into the sdk. I might be able to help with both, since one is immediately useful and we (like everyone) have deadlines to meet, and the other makes sense to me as a possible future direction that could be more widely supported. >>>>>> >>>>>> Will you or someone else be hosting and discussion about this at the Shanghai summit? I'll be there and would be happy to join and discuss. >>>>>> >>>>>> --Adam >>>>>> >>>>>> On Tue, Oct 15, 2019, 22:04 Adrian Turjak <adriant(a)catalyst.net.nz <mailto:adriant@catalyst.net.nz>> wrote: >>>>>> I tried to get a community goal to do project deletion per project, but >>>>>> we ended up deciding that a community goal wasn't ideal unless we did >>>>>> build a bulk delete API in each service: >>>>>> https://review.opendev.org/#/c/639010/ <https://review.opendev.org/#/c/639010/> >>>>>> https://etherpad.openstack.org/p/community-goal-project-deletion <https://etherpad.openstack.org/p/community-goal-project-deletion> >>>>>> https://etherpad.openstack.org/p/DEN-Deletion-of-resources <https://etherpad.openstack.org/p/DEN-Deletion-of-resources> >>>>>> https://etherpad.openstack.org/p/DEN-Train-PublicCloudWG-brainstorming <https://etherpad.openstack.org/p/DEN-Train-PublicCloudWG-brainstorming> >>>>>> >>>>>> What we decided on, but didn't get a chance to work on, was building >>>>>> into the OpenstackSDK OS-purge like functionality, as well as reporting >>>>>> functionality (of all project resources to be deleted). That way we >>>>>> could have per project per resource deletion logic, and all of that >>>>>> defined in the SDK. >>>>>> >>>>>> I was up for doing some of the work, but ended up swamped with internal >>>>>> work and just didn't drive or push for the deletion work upstream. >>>>>> >>>>>> If you want to do something useful, don't pursue OS-Purge, help us add >>>>>> that official functionality to the SDK, and then we can push for bulk >>>>>> deletion APIs in each project to make resource deletion more pleasant. >>>>>> >>>>>> I'd be happy to help with the work, and Monty on the SDK team will most >>>>>> likely be happy to as well. :) >>>>>> >>>>>> Cheers, >>>>>> Adrian >>>>>> >>>>>> On 1/10/19 11:48 am, Adam Harwell wrote: >>>>>> > I haven't seen much activity on this project in a while, and it's been >>>>>> > moved to opendev/x since the opendev migration... Who is the current >>>>>> > owner of this project? Is there anyone who actually is maintaining it, >>>>>> > or would mind if others wanted to adopt the project to move it forward? >>>>>> > >>>>>> > Thanks, >>>>>> > --Adam Harwell >>>> >>

2 1

Galera and haproxy fail; keystone error 'Lost connection to MySQL server during query'
by Albert Braden 11 Feb '20

11 Feb '20

If I run galera without haproxy and point everything to 1 server, it seems to work fine, but we want to load-balance it with haproxy. After setting up haproxy we had lots of errors in nova, neutron and keystone, but I got rid of those by setting haproxy timeout values to 10 minutes. The remaining errors are in keystone. In /var/log/keystone/keystone-wsgi-public.log I see "Loaded 2 Fernet keys" every 5 minutes and this is frequently accompanied by a mysql error 'Lost connection to MySQL server during query' I tried changing various config items in haproxy, galera and keystone, but nothing seems to help. How can I fix these errors? Errors: https://f.perl.bot/p/pbq26k https://f.perl.bot/p/1tnh78 https://f.perl.bot/p/fuxmwo Haproxy config: https://f.perl.bot/p/gu2lil Mysql timeout values: https://f.perl.bot/p/i6l7tn Keystone config (minus commented lines): https://f.perl.bot/p/o6fdht

2 3

[ptl][keystone] cmurphy afk week of February 17
by Colleen Murphy 11 Feb '20

11 Feb '20

I will be on vacation the week of February 17 and will not be computering. Kristi Nikolla (knikolla) has graciously agreed to act as stand-in PTL for keystone during that time and will chair next week's meeting. If necessary I will be reachable by email or IRC but with some delay (hours or days). Colleen

1 0

[keystone] YVR PTG planning
by Colleen Murphy 11 Feb '20

11 Feb '20

Hi all, It's time to start planning the next PTG. Unlike Shanghai, I'm tentatively assuming that we will be able to hold this one in person. I've created an etherpad to start tracking it: https://etherpad.openstack.org/p/yvr-ptg-keystone Please add your name to the etherpad if you have a hunch you'll be attending and participating in keystone-related discussions, and please also add your topic ideas to the list. Colleen

1 0