#opnfv-meeting: Octopus weekly call

Meeting summary

1. agenda bashing (ulik, 14:02:57)
   1. adding documentation system to the agenda (ulik, 14:03:28)
   2. release candidate 1 should also be on the agenda (ulik, 14:04:40)
   
   
2. additional committers (ulik, 14:05:00)
   1. Bryan Sullivan (bryan_att, 14:05:42)
   2. Uli asks whether everybody would agree to add Fatih, Iben and Jerry to the committers (ulik, 14:06:01)
   3. AGREED: Octopus team agrees that Fatih, Iben and Jerry should be committers (ulik, 14:07:30)
   4. gotomeeting link? (rprakash, 14:07:57)
   5. https://global.gotomeeting.com/join/150627173 (bryan_att, 14:08:24)
   
   
3. release candidate 1 (ulik, 14:08:57)
   1. Fatih clarifies that this Thursday, we only need to do build. (ulik, 14:10:17)
   2. This is working fine, so octopus should be fine (ulik, 14:10:44)
   3. we will verify also with BGS after this meeting. (ulik, 14:11:05)
   4. can you add me to commiter for octuopus I missed that earlier (rprakash, 14:11:14)
   5. Fuel was already running, other installers still missing (ulik, 14:11:41)
   6. Foreman, OpenSteak and others, the build scripts are still missing. (ulik, 14:12:20)
   7. Build resources are needed for different installers (fdegir, 14:13:35)
   8. Information to connect lab resources to OPNFV LF Jenkins (fdegir, 14:14:28)
   9. https://etherpad.opnfv.org/p/octopus (fdegir, 14:14:31)
   10. look for How to Connect Servers from Labs to OPNFV Jenkin (fdegir, 14:14:40)
   11. and contact aricg (fdegir, 14:15:10)
   
   
4. Triggering tests in different labs (ulik, 14:16:25)
   1. Fatih: update on triggering tests from labs (bryan_att, 14:17:14)
   2. ... some problems with internet reources being debugged for Intel lab (bryan_att, 14:17:36)
   3. ... also issues from other labs (bryan_att, 14:17:58)
   4. Proposed solution might not work for all the labs as it happened with Intel lab (fdegir, 14:18:58)
   5. uli: the more labs we try the better - all Pharos page labs should try to connect (bryan_att, 14:19:21)
   6. if this is the case for more labs, we need to get firewall openings so OPNFV Jenkins master can reach the resources in different labs (fdegir, 14:19:34)
   7. Fatih: when we get LF resources we won't have this problem (bryan_att, 14:20:14)
   8. ... it will help for R1 at least one running environment (bryan_att, 14:20:40)
   9. (?): who is working on the LF hardware? (bryan_att, 14:20:54)
   10. fatih: Constantin from LF is working on it (bryan_att, 14:21:13)
   11. aric: they are waiting on a switch; still need some topo info. but check with constantin (bryan_att, 14:21:40)
   12. uli: who will provide the info? (bryan_att, 14:21:58)
   13. peter: the suggestion was to standardize the lab setup topo (bryan_att, 14:23:03)
   14. https://wiki.opnfv.org/pharos/pharos_specification (rprakash, 14:23:22)
   15. ... last week - currently at the Pharos page the ref platform is described - public nets are missing; there are three interfaces (bryan_att, 14:23:45)
   16. ... plus lights-out net makes four (bryan_att, 14:23:59)
   17. ... another BGS meeting after this one will continue the discussion (bryan_att, 14:24:22)
   18. fatih: we need BGS and LF to address the setup so we have something for next week (bryan_att, 14:24:42)
   19. peter: we expect it to be completed soon - tests require a ref platform agreement (bryan_att, 14:25:29)
   
   
5. hardware usage (ulik, 14:26:00)
   1. uli: need more input from BGS (bryan_att, 14:26:18)
   2. fatih: for fuel we can use docker images but need input for the others e.g. bare metal or ISOs (bryan_att, 14:26:41)
   3. ?: foreman team has been working on bare metal (bryan_att, 14:27:38)
   4. ?: it would be more agile to focus on virtual environments (bryan_att, 14:28:03)
   5. uli: we need to revisit this requirement on bare metal vs virtual (bryan_att, 14:28:36)
   6. fatih: for R1 bare metal is required at least to verify its working (bryan_att, 14:29:02)
   7. some of the stack inputs from BGS to Octopus (CI) like RDO, Juju and other tracks may bnot be ISO but some artifacts like tar files (rprakash, 14:29:18)
   8. ... for the long term we can't run full CI loops for each commit and test on bare metal, but later we can do so using a virtual environment (bryan_att, 14:30:02)
   9. ... for bare metal will have to schedule based upon the installers being verified (bryan_att, 14:30:49)
   10. uli: we will revisit for R2 (bryan_att, 14:31:26)
   
   
6. document repository (ulik, 14:31:57)
   1. aric: the desired end result is PDFs - IMO not the right way, e.g. space reqs. I recommend using a markup lang and generate PDFs as an artifact (bryan_att, 14:33:06)
   2. Bryan: +1 (bryan_att, 14:33:10)
   3. fatih: some people said they don't know git etc (bryan_att, 14:33:58)
   4. bryan: we need to set an expectation that people will use the tools typical for open soruce (bryan_att, 14:34:46)
   5. uli: this must be addressed in the documentation group (bryan_att, 14:35:00)
   6. fatih: how do we handle this, e.g. go to TSC? (bryan_att, 14:35:17)
   7. uli: until the project is created, we can do it there; in the meantime in the TSC (bryan_att, 14:35:37)
   8. bryan: can Octopus make a recommendation as an agreement (bryan_att, 14:36:26)
   9. ... the documentation project should assess and support the consensus of projects rather than dictate (bryan_att, 14:37:16)
   10. ?: don't want to see word documents - rather latex (bryan_att, 14:37:46)
   11. bryan: everyone who has an opinion should voice it on the TSC and list (bryan_att, 14:39:17)
   12. fatih: can we vote on this now, here? (bryan_att, 14:39:39)
   13. For Octopus documentation: R1 use markdown, R2 latex (or something else) (fdegir, 14:40:04)
   14. ul: it would be good for Octopus to provide an input to the TSC (bryan_att, 14:41:32)
   15. to have a project voice to text tool format ? (rprakash, 14:41:39)
   16. ... anyone speaking against? (bryan_att, 14:41:42)
   17. david: agree storing a binary artifact is not good - text is better (bryan_att, 14:42:10)
   18. AGREED: to give message to tsc for a text based document format (ulik, 14:42:26)
   19. uli: we can thus use git for this (bryan_att, 14:42:54)
   20. frank: will we use markdown for R1? (bryan_att, 14:43:28)
   21. frank: all should be able to use markdown (bryan_att, 14:43:56)
   22. For Octopus, we will use markdown for R1 (fdegir, 14:44:24)
   
   
7. artifact repository (ulik, 14:45:06)
   1. bala: we can talk to TSC about whether markdown or latex should be used (bryan_att, 14:45:07)
   2. http://artifacts.opnfv.org/ (aricg, 14:45:32)
   3. fatih: aric helped setup google cloud storage for artifacts (bryan_att, 14:45:37)
   4. ... we still have some fine tuning to do (bryan_att, 14:46:03)
   5. http://artifacts.opnfv.org/ (ulik, 14:46:17)
   6. bryan: can we access control this as needed? (bryan_att, 14:46:29)
   7. aric: we can mark artifacts as private etc (bryan_att, 14:47:16)
   8. ... we will create a different bucked with specific access controls (bryan_att, 14:47:38)
   9. fatih: metadata handling is not good enough with google storage; we need to address this (bryan_att, 14:48:57)
   10. uli: if we have a R2 patch ISO needing to be deployed; this will be in google storage; can jenkins auto-store these? (bryan_att, 14:49:31)
   11. fatiih: upload/download works fine, but the problem is with metadata (bryan_att, 14:49:55)
   12. uli: the deploy script needs to find the right artifact (bryan_att, 14:50:09)
   13. fatih: we might upload our own metadata file and link it to the artifact, then extract from the storage when we want to use the artifact (bryan_att, 14:51:08)
   14. ... we can put text in both places (bryan_att, 14:51:42)
   15. ... anyone can join the discussion on #opnfv-octopus (bryan_att, 14:52:17)

Action items

1. (none)

People present (lines said)

1. bryan_att (60)
2. ulik (22)
3. fdegir (10)
4. aricg (7)
5. collabot (6)
6. rprakash (5)
7. frankbrockners (4)
8. MrSnvn (1)
9. fdegit (0)