- Release-team - mailman3.cels.anl.gov

Deadline for Feb release, and some thoughts about the release process
by Nomi Harris 21 Jan '14

21 Jan '14

Dear release team (particularly Miriam, Dan, Shane, and Tom), Michael Sneddon and I were talking today about the release process and what we need to do as we rapidly approach the February demos. The January release has not been completed; meanwhile, some developers want a release closer to the Feb cutoff as they are still fixing bugs. (I would hope that they are not adding new features at this point, though I'm pretty sure some are.) Miriam and the deploy team will obviously need to set a cutoff date for (re)deploying services, since I doubt they are planning to stay up all night the night before the DOE meeting trying to check and deploy a whole bunch of services. So, what are those dates? We probably need to reveal them to the developers or else they will assume that they can make changes until the 11th hour and get them into production for the demos. Also, Michael and I talked about release notes and how critical they are as an information source both internally and externally. We all know their importance to external users (who need to know what's changed in this release), but release notes also serve a valuable role in informing the release team about why the service needs to be (re)deployed. Is there significant new functionality, and if so, what? Are there major bug fixes? Are there changes that might affect other services upstream or downstream? There are other ways in which developers might communicate this info to the release team, but why not put it in the release notes and kill two birds with one stone? We also talked about the fact that some services have been tested and deployed and are now ready for public release EXCEPT they lack updated release notes. This is not the most efficient way to go about the release. We propose that instead of being checked last, the release notes should be sent to the release team at the time a service owner requests a release. If they don't send release notes, then their service doesn't enter the deploy pipeline until they do. Once they've sent release notes, the release team can make an informed decision about how to prioritize the release. (For example, if the release team gets a request on Feb 8 at midnight for a new release because the service has a new improved error message, they might decide that it's not worth doing a frantic last-minute redeploy of the only-slightly-changed service.) We propose the following text to be sent to developers ASAP. The only thing that needs to be filled in is the drop-dead date for requesting a release before the Feb demos. -------- The major Jan deployment/release is coming to a close. As we move towards the February demos/release, the release team is transitioning to a rolling deployment process. Instead of a single build week, production deployments will be scheduled when a service owner requests a redeploy and has met the requirements (e.g., updated release notes and tests). The FINAL date you can request a release to be sure it will be available for the February presentations is [DATE] To request a release, send an email to Miriam (landml(a)ornl.gov) and the release team (release-team(a)kbase.us) that includes: 1) name of service and/or repository to release (and the branch, if other than main) 2) updated release notes (as a git URL or attachment) documenting the major changes in this release (which will also tell the release team about the reason for the deployment). As always, your release notes should follow this format: https://trac.kbase.us/projects/kbase/wiki/ReleaseNotesStyle The release team will review the release notes to help prioritize the deployment effort and ensure that developers of dependent modules are notified; then we will proceed with deployment to production as quickly as possible. ---------- Nomi and Michael

1 0

Re: [Release-team] Next Steps in the build
by Nomi Harris 21 Jan '14

21 Jan '14

Thanks, Elizabeth. I think Miriam should be the one to say whether a service is ready for public release (and she should announce it to the release-team mailing list to make sure there are no objections). The determination about whether a service is destined for public release should have already been done (with input from Tom and the rest of the release team) before the service entered the release pipeline. Nomi On Jan 21, 2014, at 12:38 PM, Glass, Elizabeth Marie Marland <marland(a)mcs.anl.gov> wrote: > Hi all, > > I will review for column Q. Not sure if I am the one giving the OK for public release, though. > I was on the hook for the blog/social media posts. > > Elizabeth > ________________________________________ > From: Nomi Harris [nlharris(a)lbl.gov] > Sent: Tuesday, January 21, 2014 11:57 AM > To: Land, Miriam L.; Glass, Elizabeth Marie Marland > Cc: release-team(a)kbase.us > Subject: Re: [Release-team] Next Steps in the build > > Looping in Elizabeth, because I’m not sure she’s on the release-team mailing list… > > On Jan 21, 2014, at 9:22 AM, Land, Miriam L. <landml(a)ornl.gov> wrote: > >> Dan, >> >> You told me that you have a way to test that the links to the API docs were working in production. That will take care of column O. >> >> I will do column P. >> >> I believe that Elizabeth is responsible for columns Q (adequate description) and R (public release). > > Is that correct, Elizabeth? (The spreadsheet we are talking about is https://docs.google.com/a/lbl.gov/spreadsheet/ccc?key=0AriSbOyynCz1dDdJRHJB…) > > Nomi >> >> I would still like to know why workspace testing is still listed as "in process" and yet it was okayed for deployment by 'anonymous' and Shock was okayed for production without a review of documentation and release notes? It's okay to skip steps but shouldn't we at least know who did it and/or why it was done? >> >> Miriam >> >> From: <Murphy-Olson>, Dan Olson <dolson(a)mcs.anl.gov<mailto:[email protected]>> >> Date: Tuesday, January 21, 2014 11:28 AM >> To: Miriam Land <landml(a)ornl.gov<mailto:[email protected]>>, "Brettin, Thomas S." <brettin(a)cels.anl.gov<mailto:[email protected]>> >> Cc: "Gordon, India S." <Igordon(a)anl.gov<mailto:[email protected]>>, "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>> >> Subject: RE: [Release-team] please deploy id_map >> >> Miriam, >> >> No, as far as I know the spreadsheet is correct. I'm not sure who does the OPQR steps, but workspace deluxe and shock service are in there and waiting. >> >> I think we need to circle back on a lot of these that are marked "not yet". Who is handling the communication on them? >> >> >> ---- >> Daniel Murphy-Olson >> Systems Administrator >> Mathematics & Computer Science Division >> Argonne National Laboratory >> 630-252-0055 >> >> ________________________________________ >> From: Land, Miriam L. [landml(a)ornl.gov<mailto:[email protected]>] >> Sent: Tuesday, January 21, 2014 9:28 AM >> To: Brettin, Thomas S.; Murphy-Olson, Daniel E. >> Cc: Gordon, India S.; release-team(a)kbase.us<mailto:[email protected]> >> Subject: Re: [Release-team] please deploy id_map >> >> Have we completely abandoned the spreadsheet from the last build? Is there any tracking of what is being put into production? How will I know which modules need updated —help so I can send this to Lynn? >> >> Miriam >> >> From: <Brettin>, "Thomas S." <brettin(a)cels.anl.gov<mailto:[email protected]><mailto:[email protected]>> >> Date: Monday, January 20, 2014 1:07 PM >> To: Dan Olson <dolson(a)mcs.anl.gov<mailto:[email protected]><mailto:[email protected]>> >> Cc: "Gordon, India S." <Igordon(a)anl.gov<mailto:[email protected]><mailto:[email protected]>>, "release-team(a)kbase.us<mailto:[email protected]><mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]><mailto:[email protected]>> >> Subject: [Release-team] please deploy id_map >> >> To production. >> >> Please contact me with any errors. >> >> Thanks, >> Tom >> _______________________________________________ >> Release-team mailing list >> Release-team(a)lists.kbase.us<mailto:[email protected]><mailto:[email protected]> >> https://lists.kbase.us/mailman/listinfo/release-team >

1 0

Re: [Release-team] Python servers non-functional in January build
by Chia, Nicholas, Ph.D. 21 Jan '14

21 Jan '14

Awesome! Does this mean we are deployed? Nick From: Shane Canon <scanon(a)lbl.gov<mailto:[email protected]>> Date: Tue, 21 Jan 2014 14:20:05 -0800 To: "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu<mailto:[email protected]>> Cc: "Murphy-Olson, Daniel E." <dolson(a)mcs.anl.gov<mailto:[email protected]>>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu<mailto:[email protected]>>, "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>>, "Henry, Christopher S." <chrisshenry(a)gmail.com<mailto:[email protected]>>, Matthew Benedict <matthew.n.benedict(a)gmail.com<mailto:[email protected]>>, "N.L. Harris" <nlharris(a)lbl.gov<mailto:[email protected]>> Subject: Re: [Release-team] Python servers non-functional in January build Done and the tests passed. --Shane On Jan 21, 2014, at 1:56 PM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu<mailto:[email protected]>> wrote: I had a chance to talk with Dan today about the test environment and made a change to resolve our problem. Can we try the deployment of probabilistic annotation and compression based distance with the latest code in the master branch. Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org/> From: <Murphy-Olson>, "Daniel E." <dolson(a)mcs.anl.gov<mailto:[email protected]>> Date: Monday, January 20, 2014 9:44 AM To: Michael Mundy <mundy.michael(a)mayo.edu<mailto:[email protected]>>, Shane Canon <scanon(a)lbl.gov<mailto:[email protected]>> Cc: "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu<mailto:[email protected]>>, "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>>, "Henry, Christopher S." <chrisshenry(a)gmail.com<mailto:[email protected]>>, Matthew Benedict <matthew.n.benedict(a)gmail.com<mailto:[email protected]>>, Stephen Chan <sychan(a)lbl.gov<mailto:[email protected]>>, "N.L. Harris" <nlharris(a)lbl.gov<mailto:[email protected]>> Subject: RE: [Release-team] Python servers non-functional in January build t1 was used for most of the testing, and some of the testing requires the service to be running on the local host. Feel free to shut down those services and run them on the dedicated hosts if necessary. ---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055<callto:630-252-0055> ________________________________ From:[email protected]<mailto:[email protected]> [release-team-bounces(a)lists.kbase.us<mailto:[email protected]>] on behalf of Mundy, Michael B. (Mike) [Mundy.Michael(a)mayo.edu<mailto:[email protected]>] Sent: Monday, January 20, 2014 9:26 AM To: Shane Canon Cc: Chia, Nicholas, Ph.D.; release-team(a)kbase.us<mailto:[email protected]>; Henry, Christopher S.; Matthew Benedict; Stephen Chan; N.L. Harris Subject: Re: [Release-team] Python servers non-functional in January build Shane, Thanks for adding cogent. I forgot that cogent has a dependency on mpi4py (it throws annoying error messages if it is not found). I was running on t1. I'm confused by the new test environment for this build. Are all of the servers supposed to be running on t1? Or on the instances specified in /root/services? For example, the services file says that workspace_deluxe should be running on t49 at port 7058. But when I checked on t49 this morning there was nothing listening on port 7058. I'll be at Argonne Tuesday and Wednesday. Who would be the best person to work with there? Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org/> From: Shane Canon <scanon(a)lbl.gov<mailto:[email protected]>> Date: Friday, January 17, 2014 2:39 PM To: Michael Mundy <mundy.michael(a)mayo.edu<mailto:[email protected]>> Cc: "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>>, Christopher Henry <chrisshenry(a)gmail.com<mailto:[email protected]>>, Stephen Chan <sychan(a)lbl.gov<mailto:[email protected]>>, Matthew Benedict <matthew.n.benedict(a)gmail.com<mailto:[email protected]>>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu<mailto:[email protected]>>, "N.L. Harris" <nlharris(a)lbl.gov<mailto:[email protected]>> Subject: Re: [Release-team] Python servers non-functional in January build 1. I'm not sure about this. Is this on t1 or production? It could be the client tools aren't matched. ws is the new workspace which isn't "public" yet. 2. I'm seeing this too and I think it is related to going through the nginx proxy. There have been some discussions about this. 3. I added cogent on the 2014 runtime on t1. Let me know if you need more. --Shane On Jan 17, 2014, at 10:45 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu<mailto:[email protected]>> wrote: Shane, Here's what I found in my testing: 1. I used https://kbase.us/services/ws/ as the URL for the new workspace service but I'm getting a "HTTPError: HTTP Error 502: Bad Gateway" error when accessing it. 2. I'm getting a "client-tests/1_V2.fasta: error: Error creating node: request error: 411 Length Required" error when trying to upload a file to shock. The shock team had already looked at this and said it was a problem with the old shock server. The shock-client.cfg file for the ubuntu user has the server URL set to url=http://kbase.us/services/shock-api. Is that URL going to the server in the new deployment. 3. The cogent Python package isn't installed in the kbase Python installation (cogent is installed in the system Python installation). Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org/> From: Shane Canon <scanon(a)lbl.gov<mailto:[email protected]>> Date: Friday, January 17, 2014 11:01 AM To: Michael Mundy <mundy.michael(a)mayo.edu<mailto:[email protected]>> Cc: "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>>, Christopher Henry <chrisshenry(a)gmail.com<mailto:[email protected]>>, Stephen Chan <sychan(a)lbl.gov<mailto:[email protected]>>, Matthew Benedict <matthew.n.benedict(a)gmail.com<mailto:[email protected]>>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu<mailto:[email protected]>>, "N.L. Harris" <nlharris(a)lbl.gov<mailto:[email protected]>> Subject: Re: [Release-team] Python servers non-functional in January build Mike, When I originally built the runtime and deployment, I inadvertently used 2013 instead of 2014 in the path. I've created a new runtime and deployment with the correct date (/kbase/deployments/20140109-prod/ and /kbase/runtimes/20140109-prod/). That version should have Steve's changes. Can you test that out? I still need to restart a few services that were originally started with the 2013 tree. I will do that now. --Shane On Jan 17, 2014, at 8:43 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu<mailto:[email protected]>> wrote: Authenticated Python servers are currently non-functional in the January build. Steve Chan has fixed the problem with changes to the auth module with the following commits: commit 12f346a4ed8e2d30a580d85e7e7987c78568b9bb Merge: 1d98332 0eca1d4 Author: Steve Chan <sychan(a)lbl.gov<mailto:[email protected]>> Date: Mon Jan 13 15:46:37 2014 -0800 Merge branch 'master' into dev_requests-1.x Conflicts: Bio-KBase-Auth/lib/Bio/KBase/Auth.pm python-libs/biokbase/nexus/__init__.py commit 0eca1d43361b8955b0d9ce391778cd5a86561150 Author: Steve Chan <sychan(a)lbl.gov<mailto:[email protected]>> Date: Mon Jan 13 15:41:46 2014 -0800 Part of cleanup for the Nexus/requests library API drift Can we get these two commits in the build and installed on the t1 system. The Probabilistic Annotation and Compression Based Distance services cannot complete the build until this is done. Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org/> _______________________________________________ Release-team mailing list Release-team(a)lists.kbase.us<mailto:[email protected]> https://lists.kbase.us/mailman/listinfo/release-team

1 0

Re: [Release-team] Python servers non-functional in January build
by Shane Canon 21 Jan '14

21 Jan '14

Done and the tests passed. --Shane On Jan 21, 2014, at 1:56 PM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu> wrote: > I had a chance to talk with Dan today about the test environment and made a change to resolve our problem. Can we try the deployment of probabilistic annotation and compression based distance with the latest code in the master branch. > > Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu > Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org > > From: <Murphy-Olson>, "Daniel E." <dolson(a)mcs.anl.gov> > Date: Monday, January 20, 2014 9:44 AM > To: Michael Mundy <mundy.michael(a)mayo.edu>, Shane Canon <scanon(a)lbl.gov> > Cc: "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu>, "release-team(a)kbase.us" <release-team(a)kbase.us>, "Henry, Christopher S." <chrisshenry(a)gmail.com>, Matthew Benedict <matthew.n.benedict(a)gmail.com>, Stephen Chan <sychan(a)lbl.gov>, "N.L. Harris" <nlharris(a)lbl.gov> > Subject: RE: [Release-team] Python servers non-functional in January build > > t1 was used for most of the testing, and some of the testing requires the service to be running on the local host. > > Feel free to shut down those services and run them on the dedicated hosts if necessary. > > ---- > Daniel Murphy-Olson > Systems Administrator > Mathematics & Computer Science Division > Argonne National Laboratory > 630-252-0055 > From: release-team-bounces(a)lists.kbase.us [release-team-bounces(a)lists.kbase.us] on behalf of Mundy, Michael B. (Mike) [Mundy.Michael(a)mayo.edu] > Sent: Monday, January 20, 2014 9:26 AM > To: Shane Canon > Cc: Chia, Nicholas, Ph.D.; release-team(a)kbase.us; Henry, Christopher S.; Matthew Benedict; Stephen Chan; N.L. Harris > Subject: Re: [Release-team] Python servers non-functional in January build > > Shane, > > Thanks for adding cogent. I forgot that cogent has a dependency on mpi4py (it throws annoying error messages if it is not found). > > I was running on t1. I'm confused by the new test environment for this build. Are all of the servers supposed to be running on t1? Or on the instances specified in /root/services? For example, the services file says that workspace_deluxe should be running on t49 at port 7058. But when I checked on t49 this morning there was nothing listening on port 7058. > > I'll be at Argonne Tuesday and Wednesday. Who would be the best person to work with there? > > Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu > Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org > > From: Shane Canon <scanon(a)lbl.gov> > Date: Friday, January 17, 2014 2:39 PM > To: Michael Mundy <mundy.michael(a)mayo.edu> > Cc: "release-team(a)kbase.us" <release-team(a)kbase.us>, Christopher Henry <chrisshenry(a)gmail.com>, Stephen Chan <sychan(a)lbl.gov>, Matthew Benedict <matthew.n.benedict(a)gmail.com>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu>, "N.L. Harris" <nlharris(a)lbl.gov> > Subject: Re: [Release-team] Python servers non-functional in January build > > > 1. I'm not sure about this. Is this on t1 or production? It could be the client tools aren't matched. ws is the new workspace which isn't "public" yet. > > 2. I'm seeing this too and I think it is related to going through the nginx proxy. There have been some discussions about this. > > 3. I added cogent on the 2014 runtime on t1. Let me know if you need more. > > > --Shane > > > On Jan 17, 2014, at 10:45 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu> wrote: > >> Shane, >> >> Here's what I found in my testing: >> I used https://kbase.us/services/ws/ as the URL for the new workspace service but I'm getting a "HTTPError: HTTP Error 502: Bad Gateway" error when accessing it. >> I'm getting a "client-tests/1_V2.fasta: error: Error creating node: request error: 411 Length Required" error when trying to upload a file to shock. The shock team had already looked at this and said it was a problem with the old shock server. The shock-client.cfg file for the ubuntu user has the server URL set to url=http://kbase.us/services/shock-api. Is that URL going to the server in the new deployment. >> The cogent Python package isn't installed in the kbase Python installation (cogent is installed in the system Python installation). >> Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu >> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org >> >> From: Shane Canon <scanon(a)lbl.gov> >> Date: Friday, January 17, 2014 11:01 AM >> To: Michael Mundy <mundy.michael(a)mayo.edu> >> Cc: "release-team(a)kbase.us" <release-team(a)kbase.us>, Christopher Henry <chrisshenry(a)gmail.com>, Stephen Chan <sychan(a)lbl.gov>, Matthew Benedict <matthew.n.benedict(a)gmail.com>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu>, "N.L. Harris" <nlharris(a)lbl.gov> >> Subject: Re: [Release-team] Python servers non-functional in January build >> >> >> Mike, >> >> When I originally built the runtime and deployment, I inadvertently used 2013 instead of 2014 in the path. I've created a new runtime and deployment with the correct date (/kbase/deployments/20140109-prod/ and >> /kbase/runtimes/20140109-prod/). That version should have Steve's changes. Can you test that out? I still need to restart a few services that were originally started with the 2013 tree. I will do that now. >> >> >> --Shane >> >> On Jan 17, 2014, at 8:43 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu> wrote: >> >>> Authenticated Python servers are currently non-functional in the January build. Steve Chan has fixed the problem with changes to the auth module with the following commits: >>> >>> commit 12f346a4ed8e2d30a580d85e7e7987c78568b9bb >>> Merge: 1d98332 0eca1d4 >>> Author: Steve Chan <sychan(a)lbl.gov> >>> Date: Mon Jan 13 15:46:37 2014 -0800 >>> >>> Merge branch 'master' into dev_requests-1.x >>> >>> Conflicts: >>> Bio-KBase-Auth/lib/Bio/KBase/Auth.pm >>> python-libs/biokbase/nexus/__init__.py >>> >>> commit 0eca1d43361b8955b0d9ce391778cd5a86561150 >>> Author: Steve Chan <sychan(a)lbl.gov> >>> Date: Mon Jan 13 15:41:46 2014 -0800 >>> >>> Part of cleanup for the Nexus/requests library API drift >>> >>> Can we get these two commits in the build and installed on the t1 system. The Probabilistic Annotation and Compression Based Distance services cannot complete the build until this is done. >>> >>> Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu >>> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org >>> _______________________________________________ >>> Release-team mailing list >>> Release-team(a)lists.kbase.us >>> https://lists.kbase.us/mailman/listinfo/release-team >> >

1 0

Re: [Release-team] Python servers non-functional in January build
by Mundy, Michael B. (Mike) 21 Jan '14

21 Jan '14

I had a chance to talk with Dan today about the test environment and made a change to resolve our problem. Can we try the deployment of probabilistic annotation and compression based distance with the latest code in the master branch. Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org> From: <Murphy-Olson>, "Daniel E." <dolson(a)mcs.anl.gov<mailto:[email protected]>> Date: Monday, January 20, 2014 9:44 AM To: Michael Mundy <mundy.michael(a)mayo.edu<mailto:[email protected]>>, Shane Canon <scanon(a)lbl.gov<mailto:[email protected]>> Cc: "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu<mailto:[email protected]>>, "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>>, "Henry, Christopher S." <chrisshenry(a)gmail.com<mailto:[email protected]>>, Matthew Benedict <matthew.n.benedict(a)gmail.com<mailto:[email protected]>>, Stephen Chan <sychan(a)lbl.gov<mailto:[email protected]>>, "N.L. Harris" <nlharris(a)lbl.gov<mailto:[email protected]>> Subject: RE: [Release-team] Python servers non-functional in January build t1 was used for most of the testing, and some of the testing requires the service to be running on the local host. Feel free to shut down those services and run them on the dedicated hosts if necessary. ---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055<callto:630-252-0055> ________________________________ From: release-team-bounces(a)lists.kbase.us<mailto:[email protected]> [release-team-bounces(a)lists.kbase.us<mailto:[email protected]>] on behalf of Mundy, Michael B. (Mike) [Mundy.Michael(a)mayo.edu<mailto:[email protected]>] Sent: Monday, January 20, 2014 9:26 AM To: Shane Canon Cc: Chia, Nicholas, Ph.D.; release-team(a)kbase.us<mailto:[email protected]>; Henry, Christopher S.; Matthew Benedict; Stephen Chan; N.L. Harris Subject: Re: [Release-team] Python servers non-functional in January build Shane, Thanks for adding cogent. I forgot that cogent has a dependency on mpi4py (it throws annoying error messages if it is not found). I was running on t1. I'm confused by the new test environment for this build. Are all of the servers supposed to be running on t1? Or on the instances specified in /root/services? For example, the services file says that workspace_deluxe should be running on t49 at port 7058. But when I checked on t49 this morning there was nothing listening on port 7058. I'll be at Argonne Tuesday and Wednesday. Who would be the best person to work with there? Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org> From: Shane Canon <scanon(a)lbl.gov<mailto:[email protected]>> Date: Friday, January 17, 2014 2:39 PM To: Michael Mundy <mundy.michael(a)mayo.edu<mailto:[email protected]>> Cc: "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>>, Christopher Henry <chrisshenry(a)gmail.com<mailto:[email protected]>>, Stephen Chan <sychan(a)lbl.gov<mailto:[email protected]>>, Matthew Benedict <matthew.n.benedict(a)gmail.com<mailto:[email protected]>>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu<mailto:[email protected]>>, "N.L. Harris" <nlharris(a)lbl.gov<mailto:[email protected]>> Subject: Re: [Release-team] Python servers non-functional in January build 1. I'm not sure about this. Is this on t1 or production? It could be the client tools aren't matched. ws is the new workspace which isn't "public" yet. 2. I'm seeing this too and I think it is related to going through the nginx proxy. There have been some discussions about this. 3. I added cogent on the 2014 runtime on t1. Let me know if you need more. --Shane On Jan 17, 2014, at 10:45 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu<mailto:[email protected]>> wrote: Shane, Here's what I found in my testing: 1. I used https://kbase.us/services/ws/ as the URL for the new workspace service but I'm getting a "HTTPError: HTTP Error 502: Bad Gateway" error when accessing it. 2. I'm getting a "client-tests/1_V2.fasta: error: Error creating node: request error: 411 Length Required" error when trying to upload a file to shock. The shock team had already looked at this and said it was a problem with the old shock server. The shock-client.cfg file for the ubuntu user has the server URL set to url=http://kbase.us/services/shock-api. Is that URL going to the server in the new deployment. 3. The cogent Python package isn't installed in the kbase Python installation (cogent is installed in the system Python installation). Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org/> From: Shane Canon <scanon(a)lbl.gov<mailto:[email protected]>> Date: Friday, January 17, 2014 11:01 AM To: Michael Mundy <mundy.michael(a)mayo.edu<mailto:[email protected]>> Cc: "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>>, Christopher Henry <chrisshenry(a)gmail.com<mailto:[email protected]>>, Stephen Chan <sychan(a)lbl.gov<mailto:[email protected]>>, Matthew Benedict <matthew.n.benedict(a)gmail.com<mailto:[email protected]>>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu<mailto:[email protected]>>, "N.L. Harris" <nlharris(a)lbl.gov<mailto:[email protected]>> Subject: Re: [Release-team] Python servers non-functional in January build Mike, When I originally built the runtime and deployment, I inadvertently used 2013 instead of 2014 in the path. I've created a new runtime and deployment with the correct date (/kbase/deployments/20140109-prod/ and /kbase/runtimes/20140109-prod/). That version should have Steve's changes. Can you test that out? I still need to restart a few services that were originally started with the 2013 tree. I will do that now. --Shane On Jan 17, 2014, at 8:43 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu<mailto:[email protected]>> wrote: Authenticated Python servers are currently non-functional in the January build. Steve Chan has fixed the problem with changes to the auth module with the following commits: commit 12f346a4ed8e2d30a580d85e7e7987c78568b9bb Merge: 1d98332 0eca1d4 Author: Steve Chan <sychan(a)lbl.gov<mailto:[email protected]>> Date: Mon Jan 13 15:46:37 2014 -0800 Merge branch 'master' into dev_requests-1.x Conflicts: Bio-KBase-Auth/lib/Bio/KBase/Auth.pm python-libs/biokbase/nexus/__init__.py commit 0eca1d43361b8955b0d9ce391778cd5a86561150 Author: Steve Chan <sychan(a)lbl.gov<mailto:[email protected]>> Date: Mon Jan 13 15:41:46 2014 -0800 Part of cleanup for the Nexus/requests library API drift Can we get these two commits in the build and installed on the t1 system. The Probabilistic Annotation and Compression Based Distance services cannot complete the build until this is done. Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org/> _______________________________________________ Release-team mailing list Release-team(a)lists.kbase.us<mailto:[email protected]> https://lists.kbase.us/mailman/listinfo/release-team

1 0

Re: [Release-team] Next Steps in the build
by Glass, Elizabeth Marie Marland 21 Jan '14

21 Jan '14

Hi all, I will review for column Q. Not sure if I am the one giving the OK for public release, though. I was on the hook for the blog/social media posts. Elizabeth ________________________________________ From: Nomi Harris [nlharris(a)lbl.gov] Sent: Tuesday, January 21, 2014 11:57 AM To: Land, Miriam L.; Glass, Elizabeth Marie Marland Cc: release-team(a)kbase.us Subject: Re: [Release-team] Next Steps in the build Looping in Elizabeth, because I’m not sure she’s on the release-team mailing list… On Jan 21, 2014, at 9:22 AM, Land, Miriam L. <landml(a)ornl.gov> wrote: > Dan, > > You told me that you have a way to test that the links to the API docs were working in production. That will take care of column O. > > I will do column P. > > I believe that Elizabeth is responsible for columns Q (adequate description) and R (public release). Is that correct, Elizabeth? (The spreadsheet we are talking about is https://docs.google.com/a/lbl.gov/spreadsheet/ccc?key=0AriSbOyynCz1dDdJRHJB…) Nomi > > I would still like to know why workspace testing is still listed as "in process" and yet it was okayed for deployment by 'anonymous' and Shock was okayed for production without a review of documentation and release notes? It's okay to skip steps but shouldn't we at least know who did it and/or why it was done? > > Miriam > > From: <Murphy-Olson>, Dan Olson <dolson(a)mcs.anl.gov<mailto:[email protected]>> > Date: Tuesday, January 21, 2014 11:28 AM > To: Miriam Land <landml(a)ornl.gov<mailto:[email protected]>>, "Brettin, Thomas S." <brettin(a)cels.anl.gov<mailto:[email protected]>> > Cc: "Gordon, India S." <Igordon(a)anl.gov<mailto:[email protected]>>, "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>> > Subject: RE: [Release-team] please deploy id_map > > Miriam, > > No, as far as I know the spreadsheet is correct. I'm not sure who does the OPQR steps, but workspace deluxe and shock service are in there and waiting. > > I think we need to circle back on a lot of these that are marked "not yet". Who is handling the communication on them? > > > ---- > Daniel Murphy-Olson > Systems Administrator > Mathematics & Computer Science Division > Argonne National Laboratory > 630-252-0055 > > ________________________________________ > From: Land, Miriam L. [landml(a)ornl.gov<mailto:[email protected]>] > Sent: Tuesday, January 21, 2014 9:28 AM > To: Brettin, Thomas S.; Murphy-Olson, Daniel E. > Cc: Gordon, India S.; release-team(a)kbase.us<mailto:[email protected]> > Subject: Re: [Release-team] please deploy id_map > > Have we completely abandoned the spreadsheet from the last build? Is there any tracking of what is being put into production? How will I know which modules need updated —help so I can send this to Lynn? > > Miriam > > From: <Brettin>, "Thomas S." <brettin(a)cels.anl.gov<mailto:[email protected]><mailto:[email protected]>> > Date: Monday, January 20, 2014 1:07 PM > To: Dan Olson <dolson(a)mcs.anl.gov<mailto:[email protected]><mailto:[email protected]>> > Cc: "Gordon, India S." <Igordon(a)anl.gov<mailto:[email protected]><mailto:[email protected]>>, "release-team(a)kbase.us<mailto:[email protected]><mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]><mailto:[email protected]>> > Subject: [Release-team] please deploy id_map > > To production. > > Please contact me with any errors. > > Thanks, > Tom > _______________________________________________ > Release-team mailing list > Release-team(a)lists.kbase.us<mailto:[email protected]><mailto:[email protected]> > https://lists.kbase.us/mailman/listinfo/release-team

1 0

Re: [Release-team] Python servers non-functional in January build
by Mundy, Michael B. (Mike) 21 Jan '14

21 Jan '14

Thanks, that resolved the issues I was having with the Python installation. Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org> From: Shane Canon <scanon(a)lbl.gov<mailto:[email protected]>> Date: Tuesday, January 21, 2014 11:56 AM To: Michael Mundy <mundy.michael(a)mayo.edu<mailto:[email protected]>> Cc: "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>>, Christopher Henry <chrisshenry(a)gmail.com<mailto:[email protected]>>, Stephen Chan <sychan(a)lbl.gov<mailto:[email protected]>>, Matthew Benedict <matthew.n.benedict(a)gmail.com<mailto:[email protected]>>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu<mailto:[email protected]>>, "N.L. Harris" <nlharris(a)lbl.gov<mailto:[email protected]>> Subject: Re: [Release-team] Python servers non-functional in January build mpi4py should be installed now. -Shane On Jan 20, 2014, at 7:26 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu<mailto:[email protected]>> wrote: Shane, Thanks for adding cogent. I forgot that cogent has a dependency on mpi4py (it throws annoying error messages if it is not found). I was running on t1. I'm confused by the new test environment for this build. Are all of the servers supposed to be running on t1? Or on the instances specified in /root/services? For example, the services file says that workspace_deluxe should be running on t49 at port 7058. But when I checked on t49 this morning there was nothing listening on port 7058. I'll be at Argonne Tuesday and Wednesday. Who would be the best person to work with there? Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org/> From: Shane Canon <scanon(a)lbl.gov<mailto:[email protected]>> Date: Friday, January 17, 2014 2:39 PM To: Michael Mundy <mundy.michael(a)mayo.edu<mailto:[email protected]>> Cc: "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>>, Christopher Henry <chrisshenry(a)gmail.com<mailto:[email protected]>>, Stephen Chan <sychan(a)lbl.gov<mailto:[email protected]>>, Matthew Benedict <matthew.n.benedict(a)gmail.com<mailto:[email protected]>>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu<mailto:[email protected]>>, "N.L. Harris" <nlharris(a)lbl.gov<mailto:[email protected]>> Subject: Re: [Release-team] Python servers non-functional in January build 1. I'm not sure about this. Is this on t1 or production? It could be the client tools aren't matched. ws is the new workspace which isn't "public" yet. 2. I'm seeing this too and I think it is related to going through the nginx proxy. There have been some discussions about this. 3. I added cogent on the 2014 runtime on t1. Let me know if you need more. --Shane On Jan 17, 2014, at 10:45 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu<mailto:[email protected]>> wrote: Shane, Here's what I found in my testing: 1. I used https://kbase.us/services/ws/ as the URL for the new workspace service but I'm getting a "HTTPError: HTTP Error 502: Bad Gateway" error when accessing it. 2. I'm getting a "client-tests/1_V2.fasta: error: Error creating node: request error: 411 Length Required" error when trying to upload a file to shock. The shock team had already looked at this and said it was a problem with the old shock server. The shock-client.cfg file for the ubuntu user has the server URL set to url=http://kbase.us/services/shock-api. Is that URL going to the server in the new deployment. 3. The cogent Python package isn't installed in the kbase Python installation (cogent is installed in the system Python installation). Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org/> From: Shane Canon <scanon(a)lbl.gov<mailto:[email protected]>> Date: Friday, January 17, 2014 11:01 AM To: Michael Mundy <mundy.michael(a)mayo.edu<mailto:[email protected]>> Cc: "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>>, Christopher Henry <chrisshenry(a)gmail.com<mailto:[email protected]>>, Stephen Chan <sychan(a)lbl.gov<mailto:[email protected]>>, Matthew Benedict <matthew.n.benedict(a)gmail.com<mailto:[email protected]>>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu<mailto:[email protected]>>, "N.L. Harris" <nlharris(a)lbl.gov<mailto:[email protected]>> Subject: Re: [Release-team] Python servers non-functional in January build Mike, When I originally built the runtime and deployment, I inadvertently used 2013 instead of 2014 in the path. I've created a new runtime and deployment with the correct date (/kbase/deployments/20140109-prod/ and /kbase/runtimes/20140109-prod/). That version should have Steve's changes. Can you test that out? I still need to restart a few services that were originally started with the 2013 tree. I will do that now. --Shane On Jan 17, 2014, at 8:43 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu<mailto:[email protected]>> wrote: Authenticated Python servers are currently non-functional in the January build. Steve Chan has fixed the problem with changes to the auth module with the following commits: commit 12f346a4ed8e2d30a580d85e7e7987c78568b9bb Merge: 1d98332 0eca1d4 Author: Steve Chan <sychan(a)lbl.gov<mailto:[email protected]>> Date: Mon Jan 13 15:46:37 2014 -0800 Merge branch 'master' into dev_requests-1.x Conflicts: Bio-KBase-Auth/lib/Bio/KBase/Auth.pm python-libs/biokbase/nexus/__init__.py commit 0eca1d43361b8955b0d9ce391778cd5a86561150 Author: Steve Chan <sychan(a)lbl.gov<mailto:[email protected]>> Date: Mon Jan 13 15:41:46 2014 -0800 Part of cleanup for the Nexus/requests library API drift Can we get these two commits in the build and installed on the t1 system. The Probabilistic Annotation and Compression Based Distance services cannot complete the build until this is done. Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu<mailto:[email protected]> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org<http://www.mayoclinic.org/> _______________________________________________ Release-team mailing list Release-team(a)lists.kbase.us<mailto:[email protected]> https://lists.kbase.us/mailman/listinfo/release-team

1 0

Re: [Release-team] Next Steps in the build
by Land, Miriam L. 21 Jan '14

21 Jan '14

Did the Shock code change? If it didn't change, why was it part of this build? If it did change, shouldn't the release notes have been updated? That's different than just checking to see if they did change. Miriam From: <Murphy-Olson>, Dan Olson <dolson(a)mcs.anl.gov<mailto:[email protected]>> Date: Tuesday, January 21, 2014 12:44 PM To: Miriam Land <landml(a)ornl.gov<mailto:[email protected]>>, "Brettin, Thomas S." <brettin(a)cels.anl.gov<mailto:[email protected]>> Cc: "Gordon, India S." <Igordon(a)anl.gov<mailto:[email protected]>>, "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>> Subject: RE: [Release-team] Next Steps in the build Shock's documentation, and release notes hadn't changed since the last release.

1 0

Re: [Release-team] Next Steps in the build
by Nomi Harris 21 Jan '14

21 Jan '14

Looping in Elizabeth, because I’m not sure she’s on the release-team mailing list… On Jan 21, 2014, at 9:22 AM, Land, Miriam L. <landml(a)ornl.gov> wrote: > Dan, > > You told me that you have a way to test that the links to the API docs were working in production. That will take care of column O. > > I will do column P. > > I believe that Elizabeth is responsible for columns Q (adequate description) and R (public release). Is that correct, Elizabeth? (The spreadsheet we are talking about is https://docs.google.com/a/lbl.gov/spreadsheet/ccc?key=0AriSbOyynCz1dDdJRHJB…) Nomi > > I would still like to know why workspace testing is still listed as "in process" and yet it was okayed for deployment by 'anonymous' and Shock was okayed for production without a review of documentation and release notes? It's okay to skip steps but shouldn't we at least know who did it and/or why it was done? > > Miriam > > From: <Murphy-Olson>, Dan Olson <dolson(a)mcs.anl.gov<mailto:[email protected]>> > Date: Tuesday, January 21, 2014 11:28 AM > To: Miriam Land <landml(a)ornl.gov<mailto:[email protected]>>, "Brettin, Thomas S." <brettin(a)cels.anl.gov<mailto:[email protected]>> > Cc: "Gordon, India S." <Igordon(a)anl.gov<mailto:[email protected]>>, "release-team(a)kbase.us<mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]>> > Subject: RE: [Release-team] please deploy id_map > > Miriam, > > No, as far as I know the spreadsheet is correct. I'm not sure who does the OPQR steps, but workspace deluxe and shock service are in there and waiting. > > I think we need to circle back on a lot of these that are marked "not yet". Who is handling the communication on them? > > > ---- > Daniel Murphy-Olson > Systems Administrator > Mathematics & Computer Science Division > Argonne National Laboratory > 630-252-0055 > > ________________________________________ > From: Land, Miriam L. [landml(a)ornl.gov<mailto:[email protected]>] > Sent: Tuesday, January 21, 2014 9:28 AM > To: Brettin, Thomas S.; Murphy-Olson, Daniel E. > Cc: Gordon, India S.; release-team(a)kbase.us<mailto:[email protected]> > Subject: Re: [Release-team] please deploy id_map > > Have we completely abandoned the spreadsheet from the last build? Is there any tracking of what is being put into production? How will I know which modules need updated —help so I can send this to Lynn? > > Miriam > > From: <Brettin>, "Thomas S." <brettin(a)cels.anl.gov<mailto:[email protected]><mailto:[email protected]>> > Date: Monday, January 20, 2014 1:07 PM > To: Dan Olson <dolson(a)mcs.anl.gov<mailto:[email protected]><mailto:[email protected]>> > Cc: "Gordon, India S." <Igordon(a)anl.gov<mailto:[email protected]><mailto:[email protected]>>, "release-team(a)kbase.us<mailto:[email protected]><mailto:[email protected]>" <release-team(a)kbase.us<mailto:[email protected]><mailto:[email protected]>> > Subject: [Release-team] please deploy id_map > > To production. > > Please contact me with any errors. > > Thanks, > Tom > _______________________________________________ > Release-team mailing list > Release-team(a)lists.kbase.us<mailto:[email protected]><mailto:[email protected]> > https://lists.kbase.us/mailman/listinfo/release-team

1 0

Re: [Release-team] Python servers non-functional in January build
by Shane Canon 21 Jan '14

21 Jan '14

mpi4py should be installed now. -Shane On Jan 20, 2014, at 7:26 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu> wrote: > Shane, > > Thanks for adding cogent. I forgot that cogent has a dependency on mpi4py (it throws annoying error messages if it is not found). > > I was running on t1. I'm confused by the new test environment for this build. Are all of the servers supposed to be running on t1? Or on the instances specified in /root/services? For example, the services file says that workspace_deluxe should be running on t49 at port 7058. But when I checked on t49 this morning there was nothing listening on port 7058. > > I'll be at Argonne Tuesday and Wednesday. Who would be the best person to work with there? > > Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu > Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org > > From: Shane Canon <scanon(a)lbl.gov> > Date: Friday, January 17, 2014 2:39 PM > To: Michael Mundy <mundy.michael(a)mayo.edu> > Cc: "release-team(a)kbase.us" <release-team(a)kbase.us>, Christopher Henry <chrisshenry(a)gmail.com>, Stephen Chan <sychan(a)lbl.gov>, Matthew Benedict <matthew.n.benedict(a)gmail.com>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu>, "N.L. Harris" <nlharris(a)lbl.gov> > Subject: Re: [Release-team] Python servers non-functional in January build > > > 1. I'm not sure about this. Is this on t1 or production? It could be the client tools aren't matched. ws is the new workspace which isn't "public" yet. > > 2. I'm seeing this too and I think it is related to going through the nginx proxy. There have been some discussions about this. > > 3. I added cogent on the 2014 runtime on t1. Let me know if you need more. > > > --Shane > > > On Jan 17, 2014, at 10:45 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu> wrote: > >> Shane, >> >> Here's what I found in my testing: >> I used https://kbase.us/services/ws/ as the URL for the new workspace service but I'm getting a "HTTPError: HTTP Error 502: Bad Gateway" error when accessing it. >> I'm getting a "client-tests/1_V2.fasta: error: Error creating node: request error: 411 Length Required" error when trying to upload a file to shock. The shock team had already looked at this and said it was a problem with the old shock server. The shock-client.cfg file for the ubuntu user has the server URL set to url=http://kbase.us/services/shock-api. Is that URL going to the server in the new deployment. >> The cogent Python package isn't installed in the kbase Python installation (cogent is installed in the system Python installation). >> Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu >> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org >> >> From: Shane Canon <scanon(a)lbl.gov> >> Date: Friday, January 17, 2014 11:01 AM >> To: Michael Mundy <mundy.michael(a)mayo.edu> >> Cc: "release-team(a)kbase.us" <release-team(a)kbase.us>, Christopher Henry <chrisshenry(a)gmail.com>, Stephen Chan <sychan(a)lbl.gov>, Matthew Benedict <matthew.n.benedict(a)gmail.com>, "Chia, Nicholas, Ph.D." <Chia.Nicholas(a)mayo.edu>, "N.L. Harris" <nlharris(a)lbl.gov> >> Subject: Re: [Release-team] Python servers non-functional in January build >> >> >> Mike, >> >> When I originally built the runtime and deployment, I inadvertently used 2013 instead of 2014 in the path. I've created a new runtime and deployment with the correct date (/kbase/deployments/20140109-prod/ and >> /kbase/runtimes/20140109-prod/). That version should have Steve's changes. Can you test that out? I still need to restart a few services that were originally started with the 2013 tree. I will do that now. >> >> >> --Shane >> >> On Jan 17, 2014, at 8:43 AM, "Mundy, Michael B. (Mike)" <Mundy.Michael(a)mayo.edu> wrote: >> >>> Authenticated Python servers are currently non-functional in the January build. Steve Chan has fixed the problem with changes to the auth module with the following commits: >>> >>> commit 12f346a4ed8e2d30a580d85e7e7987c78568b9bb >>> Merge: 1d98332 0eca1d4 >>> Author: Steve Chan <sychan(a)lbl.gov> >>> Date: Mon Jan 13 15:46:37 2014 -0800 >>> >>> Merge branch 'master' into dev_requests-1.x >>> >>> Conflicts: >>> Bio-KBase-Auth/lib/Bio/KBase/Auth.pm >>> python-libs/biokbase/nexus/__init__.py >>> >>> commit 0eca1d43361b8955b0d9ce391778cd5a86561150 >>> Author: Steve Chan <sychan(a)lbl.gov> >>> Date: Mon Jan 13 15:41:46 2014 -0800 >>> >>> Part of cleanup for the Nexus/requests library API drift >>> >>> Can we get these two commits in the build and installed on the t1 system. The Probabilistic Annotation and Compression Based Distance services cannot complete the build until this is done. >>> >>> Mike Mundy | Senior Analyst Programmer | Bioinformatics Systems Unit | 507-266-7232 | mundy.michael(a)mayo.edu >>> Mayo Clinic | 200 First Street SW | Rochester, MN 55905 | mayoclinic.org >>> _______________________________________________ >>> Release-team mailing list >>> Release-team(a)lists.kbase.us >>> https://lists.kbase.us/mailman/listinfo/release-team >> >

1 0