[JBoss JIRA] (WFLY-8299) There is missing CS integration with core management
by Brian Stansberry (JIRA)
[ https://issues.jboss.org/browse/WFLY-8299?page=com.atlassian.jira.plugin.... ]
Brian Stansberry commented on WFLY-8299:
----------------------------------------
What is the bug here? The description is unclear to me.
> There is missing CS integration with core management
> ----------------------------------------------------
>
> Key: WFLY-8299
> URL: https://issues.jboss.org/browse/WFLY-8299
> Project: WildFly
> Issue Type: Bug
> Components: Security
> Reporter: Hynek Švábek
> Assignee: Darran Lofthouse
> Priority: Blocker
>
> Priority is BLOCKER because this issue blocks RFE verification https://issues.jboss.org/browse/EAP7-538
> There must be credential store integration with core management as is mentioned in requirements of RFE.
> *management/security-realm*
> *management/security-realm/authentication/truststore* keystore-password
> *management/security-realm/server-identity/ssl* key-password and keystore-password
> *management/security-realm/server-identity/secret*
> *management/security-realm/authentication/users*
> But *security-realm* is deprecated, these resources have this description:
> {code}
> The security-realm configuration is deprecated and may be removed or moved in future versions.
> {code}
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 2 months
[JBoss JIRA] (DROOLS-1048) Change github url from github.com/droolsjbpm/ to github.com/kiegroup/
by Geoffrey De Smet (JIRA)
[ https://issues.jboss.org/browse/DROOLS-1048?page=com.atlassian.jira.plugi... ]
Geoffrey De Smet updated DROOLS-1048:
-------------------------------------
Description:
[~mark.proctor] asked me to make an inventory of what would need to happen if we ever migrate from github.com/droolsjbpm/ to github.com/kiegroup/. Technically this issue is very simply (just change it in GitHub), but organizationally this issue is difficult and dangerous: it needs to be done with care.
https://help.github.com/articles/renaming-an-organization/
Preparation:
1) Get some experience
- Make a dummy repository organization foo_org with a dummy repository foo_repo, add some files and a few branches, fork it to your user account, add a PR on it, clone it 2 times:
-- Once from foo_org
-- Once from your user account with foo_org added as upstream ("git remote add upstream ...github.com/ foo_org/...")
- Add some local changes on both dirs.
- On GitHub, go into the repository settings, in the danger zone and move it from foo_org to foo_org_kiegroup.
- Try using it from your 2 local clones ("git pull --rebase" or "git fetch upstream"). Copy paste the error messages for later use.
-- Fix it locally without recloning, without losing the local changes, etc. Record these steps clearly in a recipe to share with everyone else later.
-- Once it's fixed, see if you can do git operations without having to specify origin for the direct foo_org clone. See also the git option "-u" - you'll likely have to add in the recipe as an extra step.
-- What happened to the open PR?
More stuff to try:
- What happens if you have a organization foo-org with repo foo-repo and you first rename the organization to bar-org and then renae the repo to bar-repo. Do local checkouts, forks and pull requests still work? So is there a double redirect?
2) (Optional - not needed now that we are moving org) Clean up our repositories in droolsjbpm. There are a few dead ones that we should not be moving.
4) Reach out
- Talk to each type of person involved:
-- R&D developer
-- QA
-- Productization
-- Product docs
-- Community
- Make an inventory of references to "github.com/droolsjbpm"
-- Do a find in path on all our files for "github.com/droolsjbpm")
-- Which webapps reference it?
--- jenkins
--- jira
--- www.openhub.net
--- gitter
--- stackoverflow
--- google forums / mailing lists
--- ...
- Set a date when you'll be renaming the org.
-- Clearly communicate on all channels to let everyone know. This includes:
--- Public mailing lists and fora: Drools, OptaPlanner, jBPM, ...
--- Internal mailing lists: sme-brms, pm lists, bsig, etc
--- IRC channel topics
--- Social media (Twitter accounts etc)
--- ...
- Clearly communicate:
-- What you'll do
-- When you'll do it
-- How strong the freeze is: Can they make local changes meanwhile? Will they lose the open PR's? How will this affect their forks and branches? Answer all these questions proactively, based on your experience. Cropped screenshots are good!
-- A recipe: What they need to do - step by step for dummies - after the merge is done.
-- Get someone (Geoffrey, Edson and Petr) to review this mail before it goes out.
- You can't overcommunicate this, but you can undercommunicate it. And there's always going to be 1 person who complains he wasn't told about it and it messes up his work.
5) Do it:
- Freeze starts, announce in capitals on every IRC channel
- Full checkout as backup (if things go to hell, pick up a phone and call us - don't rush things in trying to restore the situation in a bad way).
- *Rename the organization droolsjbpm to kiegroup* https://help.github.com/articles/renaming-an-organization/
- Make sure that no one can then create the organization droolsjbpm to break our old redirects. Contact github support for this?
- Freeze stops, announce in capitals on every IRC channel
-OLD WAY Don't do this any more:
- You can do one repository at a time, or do multiple at the same time.
-- I'd recommend to start with just 1 to get a feel for it.
- Do them in reverse order of the repository-list.txt!
- DELETE THE OLD REPOSITORY. DO NOT COPY IT, BUT MOVE IT.
-- After the move, we want anyone that's trying to pull from droolsjbpm to clearly *fail* pulling, but without losing local changes.
-- Some people will not have read your mail, they need to fail when pulling, otherwise they'll find themselves asking in 5 months why we haven't done any changes in 5 months.
- In droolsjbpm, create a repository called all_repositories_have_moved_to_github_com_kiegroup and make sure it's on the top of the list of the droolsjbpm repositories. Add a link in it's readme.adoc. This is for people who follow dead links from stackoverflow etc.
- Migrations like this should really happen during the weekend to reduce the impact on the rest of the team, ask your manager to swap a weekday for a weekend day.
OLD WAY END-
6) Clean up the fallout
- Change the pom.xml scm tag for master and 7.0.x (in all repo's that ref it)
- Go through your inventory and replace the references as far as humanly reasonable (stackoverflow is likely to be unreasonably, but openhub is not). This includes but is not limited to:
-- all code
-- openhub
-- jira
-- fisheye
-- .org websites (incl jboss.properties)
-- jboss.org
- have all developers and conbributors and QA atc change "git remote" etc to the new url's (prepare a very nice mail for this in advance with scripts that they can copy paste (windows/linux/mac).
Anyway, that's my 2 cents (from my experience when moving svn to Git and splitting up the monastical build 5 years ago). Do this right and this will be almost a non-event (5 years ago it went pretty well on the engineering side, but less on the product side). Mess this up and this can be a disaster that will antagonize a lot of people...
was:
[~mark.proctor] asked me to make an inventory of what would need to happen if we ever migrate from github.com/droolsjbpm/ to github.com/kiegroup/. Technically this issue is very simply (just change it in GitHub), but organizationally this issue is difficult and dangerous: it needs to be done with care.
Preparation:
1) Get some experience
- Make a dummy repository organization foo_org with a dummy repository foo_repo, add some files and a few branches, fork it to your user account, add a PR on it, clone it 2 times:
-- Once from foo_org
-- Once from your user account with foo_org added as upstream ("git remote add upstream ...github.com/ foo_org/...")
- Add some local changes on both dirs.
- On GitHub, go into the repository settings, in the danger zone and move it from foo_org to foo_org_kiegroup.
- Try using it from your 2 local clones ("git pull --rebase" or "git fetch upstream"). Copy paste the error messages for later use.
-- Fix it locally without recloning, without losing the local changes, etc. Record these steps clearly in a recipe to share with everyone else later.
-- Once it's fixed, see if you can do git operations without having to specify origin for the direct foo_org clone. See also the git option "-u" - you'll likely have to add in the recipe as an extra step.
-- What happened to the open PR?
More stuff to try:
- What happens if you have a organization foo-org with repo foo-repo and you first rename the organization to bar-org and then renae the repo to bar-repo. Do local checkouts, forks and pull requests still work? So is there a double redirect?
2) (Optional - not needed now that we are moving org) Clean up our repositories in droolsjbpm. There are a few dead ones that we should not be moving.
4) Reach out
- Talk to each type of person involved:
-- R&D developer
-- QA
-- Productization
-- Product docs
-- Community
- Make an inventory of references to "github.com/droolsjbpm"
-- Do a find in path on all our files for "github.com/droolsjbpm")
-- Which webapps reference it?
--- jenkins
--- jira
--- www.openhub.net
--- gitter
--- stackoverflow
--- google forums / mailing lists
--- ...
- Set a date when you'll be renaming the org.
-- Clearly communicate on all channels to let everyone know. This includes:
--- Public mailing lists and fora: Drools, OptaPlanner, jBPM, ...
--- Internal mailing lists: sme-brms, pm lists, bsig, etc
--- IRC channel topics
--- Social media (Twitter accounts etc)
--- ...
- Clearly communicate:
-- What you'll do
-- When you'll do it
-- How strong the freeze is: Can they make local changes meanwhile? Will they lose the open PR's? How will this affect their forks and branches? Answer all these questions proactively, based on your experience. Cropped screenshots are good!
-- A recipe: What they need to do - step by step for dummies - after the merge is done.
-- Get someone (Geoffrey, Edson and Petr) to review this mail before it goes out.
- You can't overcommunicate this, but you can undercommunicate it. And there's always going to be 1 person who complains he wasn't told about it and it messes up his work.
5) Do it:
- Freeze starts, announce in capitals on every IRC channel
- Full checkout as backup (if things go to hell, pick up a phone and call us - don't rush things in trying to restore the situation in a bad way).
- *Rename the organization droolsjbpm to kiegroup*
- Make sure that no one can then create the organization droolsjbpm to break our old redirects. Contact github support for this?
- Freeze stops, announce in capitals on every IRC channel
-OLD WAY Don't do this any more:
- You can do one repository at a time, or do multiple at the same time.
-- I'd recommend to start with just 1 to get a feel for it.
- Do them in reverse order of the repository-list.txt!
- DELETE THE OLD REPOSITORY. DO NOT COPY IT, BUT MOVE IT.
-- After the move, we want anyone that's trying to pull from droolsjbpm to clearly *fail* pulling, but without losing local changes.
-- Some people will not have read your mail, they need to fail when pulling, otherwise they'll find themselves asking in 5 months why we haven't done any changes in 5 months.
- In droolsjbpm, create a repository called all_repositories_have_moved_to_github_com_kiegroup and make sure it's on the top of the list of the droolsjbpm repositories. Add a link in it's readme.adoc. This is for people who follow dead links from stackoverflow etc.
- Migrations like this should really happen during the weekend to reduce the impact on the rest of the team, ask your manager to swap a weekday for a weekend day.
OLD WAY END-
6) Clean up the fallout
- Change the pom.xml scm tag for master and 7.0.x (in all repo's that ref it)
- Go through your inventory and replace the references as far as humanly reasonable (stackoverflow is likely to be unreasonably, but openhub is not). This includes but is not limited to:
-- all code
-- openhub
-- jira
-- fisheye
-- .org websites (incl jboss.properties)
-- jboss.org
- have all developers and conbributors and QA atc change "git remote" etc to the new url's (prepare a very nice mail for this in advance with scripts that they can copy paste (windows/linux/mac).
Anyway, that's my 2 cents (from my experience when moving svn to Git and splitting up the monastical build 5 years ago). Do this right and this will be almost a non-event (5 years ago it went pretty well on the engineering side, but less on the product side). Mess this up and this can be a disaster that will antagonize a lot of people...
> Change github url from github.com/droolsjbpm/ to github.com/kiegroup/
> ---------------------------------------------------------------------
>
> Key: DROOLS-1048
> URL: https://issues.jboss.org/browse/DROOLS-1048
> Project: Drools
> Issue Type: Task
> Components: build
> Reporter: Geoffrey De Smet
> Assignee: Michael Biarnes Kiefer
> Priority: Critical
>
> [~mark.proctor] asked me to make an inventory of what would need to happen if we ever migrate from github.com/droolsjbpm/ to github.com/kiegroup/. Technically this issue is very simply (just change it in GitHub), but organizationally this issue is difficult and dangerous: it needs to be done with care.
> https://help.github.com/articles/renaming-an-organization/
> Preparation:
> 1) Get some experience
> - Make a dummy repository organization foo_org with a dummy repository foo_repo, add some files and a few branches, fork it to your user account, add a PR on it, clone it 2 times:
> -- Once from foo_org
> -- Once from your user account with foo_org added as upstream ("git remote add upstream ...github.com/ foo_org/...")
> - Add some local changes on both dirs.
> - On GitHub, go into the repository settings, in the danger zone and move it from foo_org to foo_org_kiegroup.
> - Try using it from your 2 local clones ("git pull --rebase" or "git fetch upstream"). Copy paste the error messages for later use.
> -- Fix it locally without recloning, without losing the local changes, etc. Record these steps clearly in a recipe to share with everyone else later.
> -- Once it's fixed, see if you can do git operations without having to specify origin for the direct foo_org clone. See also the git option "-u" - you'll likely have to add in the recipe as an extra step.
> -- What happened to the open PR?
> More stuff to try:
> - What happens if you have a organization foo-org with repo foo-repo and you first rename the organization to bar-org and then renae the repo to bar-repo. Do local checkouts, forks and pull requests still work? So is there a double redirect?
> 2) (Optional - not needed now that we are moving org) Clean up our repositories in droolsjbpm. There are a few dead ones that we should not be moving.
> 4) Reach out
> - Talk to each type of person involved:
> -- R&D developer
> -- QA
> -- Productization
> -- Product docs
> -- Community
> - Make an inventory of references to "github.com/droolsjbpm"
> -- Do a find in path on all our files for "github.com/droolsjbpm")
> -- Which webapps reference it?
> --- jenkins
> --- jira
> --- www.openhub.net
> --- gitter
> --- stackoverflow
> --- google forums / mailing lists
> --- ...
> - Set a date when you'll be renaming the org.
> -- Clearly communicate on all channels to let everyone know. This includes:
> --- Public mailing lists and fora: Drools, OptaPlanner, jBPM, ...
> --- Internal mailing lists: sme-brms, pm lists, bsig, etc
> --- IRC channel topics
> --- Social media (Twitter accounts etc)
> --- ...
> - Clearly communicate:
> -- What you'll do
> -- When you'll do it
> -- How strong the freeze is: Can they make local changes meanwhile? Will they lose the open PR's? How will this affect their forks and branches? Answer all these questions proactively, based on your experience. Cropped screenshots are good!
> -- A recipe: What they need to do - step by step for dummies - after the merge is done.
> -- Get someone (Geoffrey, Edson and Petr) to review this mail before it goes out.
> - You can't overcommunicate this, but you can undercommunicate it. And there's always going to be 1 person who complains he wasn't told about it and it messes up his work.
> 5) Do it:
> - Freeze starts, announce in capitals on every IRC channel
> - Full checkout as backup (if things go to hell, pick up a phone and call us - don't rush things in trying to restore the situation in a bad way).
> - *Rename the organization droolsjbpm to kiegroup* https://help.github.com/articles/renaming-an-organization/
> - Make sure that no one can then create the organization droolsjbpm to break our old redirects. Contact github support for this?
> - Freeze stops, announce in capitals on every IRC channel
> -OLD WAY Don't do this any more:
> - You can do one repository at a time, or do multiple at the same time.
> -- I'd recommend to start with just 1 to get a feel for it.
> - Do them in reverse order of the repository-list.txt!
> - DELETE THE OLD REPOSITORY. DO NOT COPY IT, BUT MOVE IT.
> -- After the move, we want anyone that's trying to pull from droolsjbpm to clearly *fail* pulling, but without losing local changes.
> -- Some people will not have read your mail, they need to fail when pulling, otherwise they'll find themselves asking in 5 months why we haven't done any changes in 5 months.
> - In droolsjbpm, create a repository called all_repositories_have_moved_to_github_com_kiegroup and make sure it's on the top of the list of the droolsjbpm repositories. Add a link in it's readme.adoc. This is for people who follow dead links from stackoverflow etc.
> - Migrations like this should really happen during the weekend to reduce the impact on the rest of the team, ask your manager to swap a weekday for a weekend day.
> OLD WAY END-
> 6) Clean up the fallout
> - Change the pom.xml scm tag for master and 7.0.x (in all repo's that ref it)
> - Go through your inventory and replace the references as far as humanly reasonable (stackoverflow is likely to be unreasonably, but openhub is not). This includes but is not limited to:
> -- all code
> -- openhub
> -- jira
> -- fisheye
> -- .org websites (incl jboss.properties)
> -- jboss.org
> - have all developers and conbributors and QA atc change "git remote" etc to the new url's (prepare a very nice mail for this in advance with scripts that they can copy paste (windows/linux/mac).
> Anyway, that's my 2 cents (from my experience when moving svn to Git and splitting up the monastical build 5 years ago). Do this right and this will be almost a non-event (5 years ago it went pretty well on the engineering side, but less on the product side). Mess this up and this can be a disaster that will antagonize a lot of people...
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 2 months
[JBoss JIRA] (DROOLS-1048) Change github url from github.com/droolsjbpm/ to github.com/kiegroup/
by Geoffrey De Smet (JIRA)
[ https://issues.jboss.org/browse/DROOLS-1048?page=com.atlassian.jira.plugi... ]
Geoffrey De Smet edited comment on DROOLS-1048 at 3/6/17 9:53 AM:
------------------------------------------------------------------
Separate issue to be done later: *All repository names must match the artifactId of the root pom.*
- repo droolsjbpm-knowledge must be renamed to ...
- -pom drools-multiproject must be renamed to drools (to match rep drools)-
etc
This is important for automated scripts such as xmvn (which builds RPM's), as well as to adhere to maven standards (which are relied upon by maven site etc).
was (Author: ge0ffrey):
Extra requirement: *All repository names must match the artifactId of the root pom.*
- repo droolsjbpm-knowledge must be renamed to ...
- -pom drools-multiproject must be renamed to drools (to match rep drools)-
etc
This is important for automated scripts such as xmvn (which builds RPM's), as well as to adhere to maven standards (which are relied upon by maven site etc).
> Change github url from github.com/droolsjbpm/ to github.com/kiegroup/
> ---------------------------------------------------------------------
>
> Key: DROOLS-1048
> URL: https://issues.jboss.org/browse/DROOLS-1048
> Project: Drools
> Issue Type: Task
> Components: build
> Reporter: Geoffrey De Smet
> Assignee: Michael Biarnes Kiefer
> Priority: Critical
>
> [~mark.proctor] asked me to make an inventory of what would need to happen if we ever migrate from github.com/droolsjbpm/ to github.com/kiegroup/. Technically this issue is very simply (just change it in GitHub), but organizationally this issue is difficult and dangerous: it needs to be done with care.
> Preparation:
> 1) Get some experience
> - Make a dummy repository organization foo_org with a dummy repository foo_repo, add some files and a few branches, fork it to your user account, add a PR on it, clone it 2 times:
> -- Once from foo_org
> -- Once from your user account with foo_org added as upstream ("git remote add upstream ...github.com/ foo_org/...")
> - Add some local changes on both dirs.
> - On GitHub, go into the repository settings, in the danger zone and move it from foo_org to foo_org_kiegroup.
> - Try using it from your 2 local clones ("git pull --rebase" or "git fetch upstream"). Copy paste the error messages for later use.
> -- Fix it locally without recloning, without losing the local changes, etc. Record these steps clearly in a recipe to share with everyone else later.
> -- Once it's fixed, see if you can do git operations without having to specify origin for the direct foo_org clone. See also the git option "-u" - you'll likely have to add in the recipe as an extra step.
> -- What happened to the open PR?
> More stuff to try:
> - What happens if you have a organization foo-org with repo foo-repo and you first rename the organization to bar-org and then renae the repo to bar-repo. Do local checkouts, forks and pull requests still work? So is there a double redirect?
> 2) (Optional - not needed now that we are moving org) Clean up our repositories in droolsjbpm. There are a few dead ones that we should not be moving.
> 4) Reach out
> - Talk to each type of person involved:
> -- R&D developer
> -- QA
> -- Productization
> -- Product docs
> -- Community
> - Make an inventory of references to "github.com/droolsjbpm"
> -- Do a find in path on all our files for "github.com/droolsjbpm")
> -- Which webapps reference it?
> --- jenkins
> --- jira
> --- www.openhub.net
> --- gitter
> --- stackoverflow
> --- google forums / mailing lists
> --- ...
> - Set a date when you'll be renaming the org.
> -- Clearly communicate on all channels to let everyone know. This includes:
> --- Public mailing lists and fora: Drools, OptaPlanner, jBPM, ...
> --- Internal mailing lists: sme-brms, pm lists, bsig, etc
> --- IRC channel topics
> --- Social media (Twitter accounts etc)
> --- ...
> - Clearly communicate:
> -- What you'll do
> -- When you'll do it
> -- How strong the freeze is: Can they make local changes meanwhile? Will they lose the open PR's? How will this affect their forks and branches? Answer all these questions proactively, based on your experience. Cropped screenshots are good!
> -- A recipe: What they need to do - step by step for dummies - after the merge is done.
> -- Get someone (Geoffrey, Edson and Petr) to review this mail before it goes out.
> - You can't overcommunicate this, but you can undercommunicate it. And there's always going to be 1 person who complains he wasn't told about it and it messes up his work.
> 5) Do it:
> - Freeze starts, announce in capitals on every IRC channel
> - Full checkout as backup (if things go to hell, pick up a phone and call us - don't rush things in trying to restore the situation in a bad way).
> - *Rename the organization droolsjbpm to kiegroup*
> - Make sure that no one can then create the organization droolsjbpm to break our old redirects. Contact github support for this?
> - Freeze stops, announce in capitals on every IRC channel
> -OLD WAY Don't do this any more:
> - You can do one repository at a time, or do multiple at the same time.
> -- I'd recommend to start with just 1 to get a feel for it.
> - Do them in reverse order of the repository-list.txt!
> - DELETE THE OLD REPOSITORY. DO NOT COPY IT, BUT MOVE IT.
> -- After the move, we want anyone that's trying to pull from droolsjbpm to clearly *fail* pulling, but without losing local changes.
> -- Some people will not have read your mail, they need to fail when pulling, otherwise they'll find themselves asking in 5 months why we haven't done any changes in 5 months.
> - In droolsjbpm, create a repository called all_repositories_have_moved_to_github_com_kiegroup and make sure it's on the top of the list of the droolsjbpm repositories. Add a link in it's readme.adoc. This is for people who follow dead links from stackoverflow etc.
> - Migrations like this should really happen during the weekend to reduce the impact on the rest of the team, ask your manager to swap a weekday for a weekend day.
> OLD WAY END-
> 6) Clean up the fallout
> - Change the pom.xml scm tag for master and 7.0.x (in all repo's that ref it)
> - Go through your inventory and replace the references as far as humanly reasonable (stackoverflow is likely to be unreasonably, but openhub is not). This includes but is not limited to:
> -- all code
> -- openhub
> -- jira
> -- fisheye
> -- .org websites (incl jboss.properties)
> -- jboss.org
> - have all developers and conbributors and QA atc change "git remote" etc to the new url's (prepare a very nice mail for this in advance with scripts that they can copy paste (windows/linux/mac).
> Anyway, that's my 2 cents (from my experience when moving svn to Git and splitting up the monastical build 5 years ago). Do this right and this will be almost a non-event (5 years ago it went pretty well on the engineering side, but less on the product side). Mess this up and this can be a disaster that will antagonize a lot of people...
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 2 months
[JBoss JIRA] (DROOLS-1048) Change github url from github.com/droolsjbpm/ to github.com/kiegroup/
by Geoffrey De Smet (JIRA)
[ https://issues.jboss.org/browse/DROOLS-1048?page=com.atlassian.jira.plugi... ]
Geoffrey De Smet updated DROOLS-1048:
-------------------------------------
Description:
[~mark.proctor] asked me to make an inventory of what would need to happen if we ever migrate from github.com/droolsjbpm/ to github.com/kiegroup/. Technically this issue is very simply (just change it in GitHub), but organizationally this issue is difficult and dangerous: it needs to be done with care.
Preparation:
1) Get some experience
- Make a dummy repository organization foo_org with a dummy repository foo_repo, add some files and a few branches, fork it to your user account, add a PR on it, clone it 2 times:
-- Once from foo_org
-- Once from your user account with foo_org added as upstream ("git remote add upstream ...github.com/ foo_org/...")
- Add some local changes on both dirs.
- On GitHub, go into the repository settings, in the danger zone and move it from foo_org to foo_org_kiegroup.
- Try using it from your 2 local clones ("git pull --rebase" or "git fetch upstream"). Copy paste the error messages for later use.
-- Fix it locally without recloning, without losing the local changes, etc. Record these steps clearly in a recipe to share with everyone else later.
-- Once it's fixed, see if you can do git operations without having to specify origin for the direct foo_org clone. See also the git option "-u" - you'll likely have to add in the recipe as an extra step.
-- What happened to the open PR?
More stuff to try:
- What happens if you have a organization foo-org with repo foo-repo and you first rename the organization to bar-org and then renae the repo to bar-repo. Do local checkouts, forks and pull requests still work? So is there a double redirect?
2) (Optional - not needed now that we are moving org) Clean up our repositories in droolsjbpm. There are a few dead ones that we should not be moving.
4) Reach out
- Talk to each type of person involved:
-- R&D developer
-- QA
-- Productization
-- Product docs
-- Community
- Make an inventory of references to "github.com/droolsjbpm"
-- Do a find in path on all our files for "github.com/droolsjbpm")
-- Which webapps reference it?
--- jenkins
--- jira
--- www.openhub.net
--- gitter
--- stackoverflow
--- google forums / mailing lists
--- ...
- Set a date when you'll be renaming the org.
-- Clearly communicate on all channels to let everyone know. This includes:
--- Public mailing lists and fora: Drools, OptaPlanner, jBPM, ...
--- Internal mailing lists: sme-brms, pm lists, bsig, etc
--- IRC channel topics
--- Social media (Twitter accounts etc)
--- ...
- Clearly communicate:
-- What you'll do
-- When you'll do it
-- How strong the freeze is: Can they make local changes meanwhile? Will they lose the open PR's? How will this affect their forks and branches? Answer all these questions proactively, based on your experience. Cropped screenshots are good!
-- A recipe: What they need to do - step by step for dummies - after the merge is done.
-- Get someone (Geoffrey, Edson and Petr) to review this mail before it goes out.
- You can't overcommunicate this, but you can undercommunicate it. And there's always going to be 1 person who complains he wasn't told about it and it messes up his work.
5) Do it:
- Freeze starts, announce in capitals on every IRC channel
- Full checkout as backup (if things go to hell, pick up a phone and call us - don't rush things in trying to restore the situation in a bad way).
- *Rename the organization droolsjbpm to kiegroup*
- Make sure that no one can then create the organization droolsjbpm to break our old redirects. Contact github support for this?
- Freeze stops, announce in capitals on every IRC channel
-OLD WAY Don't do this any more:
- You can do one repository at a time, or do multiple at the same time.
-- I'd recommend to start with just 1 to get a feel for it.
- Do them in reverse order of the repository-list.txt!
- DELETE THE OLD REPOSITORY. DO NOT COPY IT, BUT MOVE IT.
-- After the move, we want anyone that's trying to pull from droolsjbpm to clearly *fail* pulling, but without losing local changes.
-- Some people will not have read your mail, they need to fail when pulling, otherwise they'll find themselves asking in 5 months why we haven't done any changes in 5 months.
- In droolsjbpm, create a repository called all_repositories_have_moved_to_github_com_kiegroup and make sure it's on the top of the list of the droolsjbpm repositories. Add a link in it's readme.adoc. This is for people who follow dead links from stackoverflow etc.
- Migrations like this should really happen during the weekend to reduce the impact on the rest of the team, ask your manager to swap a weekday for a weekend day.
OLD WAY END-
6) Clean up the fallout
- Change the pom.xml scm tag for master and 7.0.x (in all repo's that ref it)
- Go through your inventory and replace the references as far as humanly reasonable (stackoverflow is likely to be unreasonably, but openhub is not). This includes but is not limited to:
-- all code
-- openhub
-- jira
-- fisheye
-- .org websites (incl jboss.properties)
-- jboss.org
- have all developers and conbributors and QA atc change "git remote" etc to the new url's (prepare a very nice mail for this in advance with scripts that they can copy paste (windows/linux/mac).
Anyway, that's my 2 cents (from my experience when moving svn to Git and splitting up the monastical build 5 years ago). Do this right and this will be almost a non-event (5 years ago it went pretty well on the engineering side, but less on the product side). Mess this up and this can be a disaster that will antagonize a lot of people...
was:
[~mark.proctor] asked me to make an inventory of what would need to happen if we ever migrate from github.com/droolsjbpm/ to github.com/kiegroup/. Technically this issue is very simply (just change it in GitHub), but organizationally this issue is difficult and dangerous: it needs to be done with care.
Preparation:
1) Get some experience
- Make a dummy repository in droolsjbpm, add some files and a few branches, fork it to your user account, add a PR on it, clone it 2 times:
-- Once from droolsjbpm
-- Once from your user account with droolsjbpm addes as upstream ("git remote add upstream ...github.com/droolsjbpm/...")
- Add some local changes on both dirs.
- On GitHub, go into the repository settings, in the danger zone and move it from droolsjbpm to kiegroup.
- Try using it from your 2 local clones ("git pull --rebase" or "git fetch upstream"). Copy paste the error messages for later use.
-- Fix it locally without recloning, without losing the local changes, etc. Record these steps clearly in a recipe to share with everyone else later.
-- Once it's fixed, see if you can do git operations without having to specify origin for the direct droolsjbpm clone. See also the git option "-u" - you'll likely have to add in the recipe as an extra step.
-- What happened to the open PR?
2) Clean up our repositories in droolsjbpm. There are a few dead ones that we should not be moving.
3) Set up roles and users of github.com/kiegroup.
4) Reach out
- Talk to each type of person involved:
-- R&D developer
-- QA
-- Productization
-- Product docs
-- Community
- Make an inventory of references to "github.com/droolsjbpm"
-- Do a find in path on all our files for "github.com/droolsjbpm")
-- Which webapps reference it?
--- jenkins
--- jira
--- www.openhub.net
--- gitter
--- stackoverflow
--- google forums / mailing lists
--- ...
- Set a date when you'll be moving the first repository.
-- Clearly communicate on all channels to let everyone know. This includes:
--- Public mailing lists and fora: Drools, OptaPlanner, jBPM, ...
--- Internal mailing lists: sme-brms, pm lists, bsig, etc
--- IRC channel topics
--- Social media (Twitter accounts etc)
--- ...
- Clearly communicate:
-- What you'll do
-- When you'll do it
-- How strong the freeze is: Can they make local changes meanwhile? Will they lose the open PR's? How will this affect their forks and branches? Answer all these questions proactively, based on your experience. Cropped screenshots are good!
-- A recipe: What they need to do - step by step for dummies - after the merge is done.
-- Get someone to review this mail before it goes out.
- You can't overcommunicate this, but you can undercommunicate it. And there's always going to be 1 person who complains he wasn't told about it and it messes up his work.
5) Do it:
- Rename the organization droolsjbpm to kiegroup
- Make sure that no one can then create the organization droolsjbpm to break our old redirects. Contact github support for this?
-OLD WAY
- You can do one repository at a time, or do multiple at the same time.
-- I'd recommend to start with just 1 to get a feel for it.
- Do them in reverse order of the repository-list.txt!
- DELETE THE OLD REPOSITORY. DO NOT COPY IT, BUT MOVE IT.
-- After the move, we want anyone that's trying to pull from droolsjbpm to clearly *fail* pulling, but without losing local changes.
-- Some people will not have read your mail, they need to fail when pulling, otherwise they'll find themselves asking in 5 months why we haven't done any changes in 5 months.
- In droolsjbpm, create a repository called all_repositories_have_moved_to_github_com_kiegroup and make sure it's on the top of the list of the droolsjbpm repositories. Add a link in it's readme.adoc. This is for people who follow dead links from stackoverflow etc.
- Migrations like this should really happen during the weekend to reduce the impact on the rest of the team, ask your manager to swap a weekday for a weekend day.-
6) Clean up the fallout
- Go through your inventory and replace the references as far as humanly reasonable (stackoverflow is likely to be unreasonably, but openhub is not).
Anyway, that's my 2 cents (from my experience when moving svn to Git and splitting up the monastical build 5 years ago). Do this right and this will be almost a non-event (5 years ago it went pretty well on the engineering side, but less on the product side). Mess this up and this can be a disaster that will antagonize a lot of people...
> Change github url from github.com/droolsjbpm/ to github.com/kiegroup/
> ---------------------------------------------------------------------
>
> Key: DROOLS-1048
> URL: https://issues.jboss.org/browse/DROOLS-1048
> Project: Drools
> Issue Type: Task
> Components: build
> Reporter: Geoffrey De Smet
> Assignee: Michael Biarnes Kiefer
> Priority: Critical
>
> [~mark.proctor] asked me to make an inventory of what would need to happen if we ever migrate from github.com/droolsjbpm/ to github.com/kiegroup/. Technically this issue is very simply (just change it in GitHub), but organizationally this issue is difficult and dangerous: it needs to be done with care.
> Preparation:
> 1) Get some experience
> - Make a dummy repository organization foo_org with a dummy repository foo_repo, add some files and a few branches, fork it to your user account, add a PR on it, clone it 2 times:
> -- Once from foo_org
> -- Once from your user account with foo_org added as upstream ("git remote add upstream ...github.com/ foo_org/...")
> - Add some local changes on both dirs.
> - On GitHub, go into the repository settings, in the danger zone and move it from foo_org to foo_org_kiegroup.
> - Try using it from your 2 local clones ("git pull --rebase" or "git fetch upstream"). Copy paste the error messages for later use.
> -- Fix it locally without recloning, without losing the local changes, etc. Record these steps clearly in a recipe to share with everyone else later.
> -- Once it's fixed, see if you can do git operations without having to specify origin for the direct foo_org clone. See also the git option "-u" - you'll likely have to add in the recipe as an extra step.
> -- What happened to the open PR?
> More stuff to try:
> - What happens if you have a organization foo-org with repo foo-repo and you first rename the organization to bar-org and then renae the repo to bar-repo. Do local checkouts, forks and pull requests still work? So is there a double redirect?
> 2) (Optional - not needed now that we are moving org) Clean up our repositories in droolsjbpm. There are a few dead ones that we should not be moving.
> 4) Reach out
> - Talk to each type of person involved:
> -- R&D developer
> -- QA
> -- Productization
> -- Product docs
> -- Community
> - Make an inventory of references to "github.com/droolsjbpm"
> -- Do a find in path on all our files for "github.com/droolsjbpm")
> -- Which webapps reference it?
> --- jenkins
> --- jira
> --- www.openhub.net
> --- gitter
> --- stackoverflow
> --- google forums / mailing lists
> --- ...
> - Set a date when you'll be renaming the org.
> -- Clearly communicate on all channels to let everyone know. This includes:
> --- Public mailing lists and fora: Drools, OptaPlanner, jBPM, ...
> --- Internal mailing lists: sme-brms, pm lists, bsig, etc
> --- IRC channel topics
> --- Social media (Twitter accounts etc)
> --- ...
> - Clearly communicate:
> -- What you'll do
> -- When you'll do it
> -- How strong the freeze is: Can they make local changes meanwhile? Will they lose the open PR's? How will this affect their forks and branches? Answer all these questions proactively, based on your experience. Cropped screenshots are good!
> -- A recipe: What they need to do - step by step for dummies - after the merge is done.
> -- Get someone (Geoffrey, Edson and Petr) to review this mail before it goes out.
> - You can't overcommunicate this, but you can undercommunicate it. And there's always going to be 1 person who complains he wasn't told about it and it messes up his work.
> 5) Do it:
> - Freeze starts, announce in capitals on every IRC channel
> - Full checkout as backup (if things go to hell, pick up a phone and call us - don't rush things in trying to restore the situation in a bad way).
> - *Rename the organization droolsjbpm to kiegroup*
> - Make sure that no one can then create the organization droolsjbpm to break our old redirects. Contact github support for this?
> - Freeze stops, announce in capitals on every IRC channel
> -OLD WAY Don't do this any more:
> - You can do one repository at a time, or do multiple at the same time.
> -- I'd recommend to start with just 1 to get a feel for it.
> - Do them in reverse order of the repository-list.txt!
> - DELETE THE OLD REPOSITORY. DO NOT COPY IT, BUT MOVE IT.
> -- After the move, we want anyone that's trying to pull from droolsjbpm to clearly *fail* pulling, but without losing local changes.
> -- Some people will not have read your mail, they need to fail when pulling, otherwise they'll find themselves asking in 5 months why we haven't done any changes in 5 months.
> - In droolsjbpm, create a repository called all_repositories_have_moved_to_github_com_kiegroup and make sure it's on the top of the list of the droolsjbpm repositories. Add a link in it's readme.adoc. This is for people who follow dead links from stackoverflow etc.
> - Migrations like this should really happen during the weekend to reduce the impact on the rest of the team, ask your manager to swap a weekday for a weekend day.
> OLD WAY END-
> 6) Clean up the fallout
> - Change the pom.xml scm tag for master and 7.0.x (in all repo's that ref it)
> - Go through your inventory and replace the references as far as humanly reasonable (stackoverflow is likely to be unreasonably, but openhub is not). This includes but is not limited to:
> -- all code
> -- openhub
> -- jira
> -- fisheye
> -- .org websites (incl jboss.properties)
> -- jboss.org
> - have all developers and conbributors and QA atc change "git remote" etc to the new url's (prepare a very nice mail for this in advance with scripts that they can copy paste (windows/linux/mac).
> Anyway, that's my 2 cents (from my experience when moving svn to Git and splitting up the monastical build 5 years ago). Do this right and this will be almost a non-event (5 years ago it went pretty well on the engineering side, but less on the product side). Mess this up and this can be a disaster that will antagonize a lot of people...
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 2 months
[JBoss JIRA] (WFLY-8300) standalone-picketlink.xml boots with lots of errors
by Darran Lofthouse (JIRA)
[ https://issues.jboss.org/browse/WFLY-8300?page=com.atlassian.jira.plugin.... ]
Darran Lofthouse commented on WFLY-8300:
----------------------------------------
This issue is a good illustration of why we don't have a standalone-elytron.xml anymore and why I would be reluctant to add it back. We have these specialist configurations but they are not used within the testsuite so although we probably broke this a long time ago it is only just discovered now.
> standalone-picketlink.xml boots with lots of errors
> ---------------------------------------------------
>
> Key: WFLY-8300
> URL: https://issues.jboss.org/browse/WFLY-8300
> Project: WildFly
> Issue Type: Bug
> Components: Security
> Reporter: Tomaz Cerar
> Assignee: Darran Lofthouse
>
> If you try to run standalone-picketlink.xml server configuration, it's boot fails with lots of problems. They look related to elytron changes, but I am not sure.
> {noformat}
> 15:27:36,476 ERROR [org.jboss.as.controller.management-operation] (Controller Boot Thread) WFLYCTL0013: Operation ("add") failed - address: ([
> ("subsystem" => "picketlink-identity-management"),
> ("partition-manager" => "jpa.emf.modules.partition.manager")
> ]) - failure description: {
> "WFLYCTL0080: Failed services" => {"jboss.picketlink-identity-management.\"jpa.emf.modules.partition.manager\".\"jpa.config\".jpa-store" => "org.jboss.msc.service.StartException in service jb
> oss.picketlink-identity-management.\"jpa.emf.modules.partition.manager\".\"jpa.config\".jpa-store: Failed to start service
> Caused by: org.picketlink.idm.config.SecurityConfigurationException: WFLYPL0050: Entities module not found [my.module]."},
> "WFLYCTL0412: Required services that are not installed:" => ["jboss.picketlink-identity-management.\"jpa.emf.modules.partition.manager\".\"jpa.config\".jpa-store"]
> }
> 15:27:36,477 ERROR [org.jboss.as.controller.management-operation] (Controller Boot Thread) WFLYCTL0013: Operation ("add") failed - address: ([
> ("subsystem" => "undertow"),
> ("server" => "default-server"),
> ("host" => "default-host"),
> ("setting" => "http-invoker")
> ]) - failure description: {
> "WFLYCTL0412: Required services that are not installed:" => ["org.wildfly.security.http-authentication-factory.application-http-authentication"],
> "WFLYCTL0180: Services with missing/unavailable dependencies" => ["jboss.undertow.server.default-server.default-host.http-invoker is missing [org.wildfly.security.http-authentication-factory.
> application-http-authentication]"]
> }
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 2 months
[JBoss JIRA] (WFLY-8300) standalone-picketlink.xml boots with lots of errors
by Tomaz Cerar (JIRA)
Tomaz Cerar created WFLY-8300:
---------------------------------
Summary: standalone-picketlink.xml boots with lots of errors
Key: WFLY-8300
URL: https://issues.jboss.org/browse/WFLY-8300
Project: WildFly
Issue Type: Bug
Components: Security
Reporter: Tomaz Cerar
Assignee: Darran Lofthouse
If you try to run standalone-picketlink.xml server configuration, it's boot fails with lots of problems. They look related to elytron changes, but I am not sure.
{noformat}
15:27:36,476 ERROR [org.jboss.as.controller.management-operation] (Controller Boot Thread) WFLYCTL0013: Operation ("add") failed - address: ([
("subsystem" => "picketlink-identity-management"),
("partition-manager" => "jpa.emf.modules.partition.manager")
]) - failure description: {
"WFLYCTL0080: Failed services" => {"jboss.picketlink-identity-management.\"jpa.emf.modules.partition.manager\".\"jpa.config\".jpa-store" => "org.jboss.msc.service.StartException in service jb
oss.picketlink-identity-management.\"jpa.emf.modules.partition.manager\".\"jpa.config\".jpa-store: Failed to start service
Caused by: org.picketlink.idm.config.SecurityConfigurationException: WFLYPL0050: Entities module not found [my.module]."},
"WFLYCTL0412: Required services that are not installed:" => ["jboss.picketlink-identity-management.\"jpa.emf.modules.partition.manager\".\"jpa.config\".jpa-store"]
}
15:27:36,477 ERROR [org.jboss.as.controller.management-operation] (Controller Boot Thread) WFLYCTL0013: Operation ("add") failed - address: ([
("subsystem" => "undertow"),
("server" => "default-server"),
("host" => "default-host"),
("setting" => "http-invoker")
]) - failure description: {
"WFLYCTL0412: Required services that are not installed:" => ["org.wildfly.security.http-authentication-factory.application-http-authentication"],
"WFLYCTL0180: Services with missing/unavailable dependencies" => ["jboss.undertow.server.default-server.default-host.http-invoker is missing [org.wildfly.security.http-authentication-factory.
application-http-authentication]"]
}
{noformat}
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 2 months
[JBoss JIRA] (DROOLS-1048) Change github url from github.com/droolsjbpm/ to github.com/kiegroup/
by Geoffrey De Smet (JIRA)
[ https://issues.jboss.org/browse/DROOLS-1048?page=com.atlassian.jira.plugi... ]
Geoffrey De Smet updated DROOLS-1048:
-------------------------------------
Description:
[~mark.proctor] asked me to make an inventory of what would need to happen if we ever migrate from github.com/droolsjbpm/ to github.com/kiegroup/. Technically this issue is very simply (just change it in GitHub), but organizationally this issue is difficult and dangerous: it needs to be done with care.
Preparation:
1) Get some experience
- Make a dummy repository in droolsjbpm, add some files and a few branches, fork it to your user account, add a PR on it, clone it 2 times:
-- Once from droolsjbpm
-- Once from your user account with droolsjbpm addes as upstream ("git remote add upstream ...github.com/droolsjbpm/...")
- Add some local changes on both dirs.
- On GitHub, go into the repository settings, in the danger zone and move it from droolsjbpm to kiegroup.
- Try using it from your 2 local clones ("git pull --rebase" or "git fetch upstream"). Copy paste the error messages for later use.
-- Fix it locally without recloning, without losing the local changes, etc. Record these steps clearly in a recipe to share with everyone else later.
-- Once it's fixed, see if you can do git operations without having to specify origin for the direct droolsjbpm clone. See also the git option "-u" - you'll likely have to add in the recipe as an extra step.
-- What happened to the open PR?
2) Clean up our repositories in droolsjbpm. There are a few dead ones that we should not be moving.
3) Set up roles and users of github.com/kiegroup.
4) Reach out
- Talk to each type of person involved:
-- R&D developer
-- QA
-- Productization
-- Product docs
-- Community
- Make an inventory of references to "github.com/droolsjbpm"
-- Do a find in path on all our files for "github.com/droolsjbpm")
-- Which webapps reference it?
--- jenkins
--- jira
--- www.openhub.net
--- gitter
--- stackoverflow
--- google forums / mailing lists
--- ...
- Set a date when you'll be moving the first repository.
-- Clearly communicate on all channels to let everyone know. This includes:
--- Public mailing lists and fora: Drools, OptaPlanner, jBPM, ...
--- Internal mailing lists: sme-brms, pm lists, bsig, etc
--- IRC channel topics
--- Social media (Twitter accounts etc)
--- ...
- Clearly communicate:
-- What you'll do
-- When you'll do it
-- How strong the freeze is: Can they make local changes meanwhile? Will they lose the open PR's? How will this affect their forks and branches? Answer all these questions proactively, based on your experience. Cropped screenshots are good!
-- A recipe: What they need to do - step by step for dummies - after the merge is done.
-- Get someone to review this mail before it goes out.
- You can't overcommunicate this, but you can undercommunicate it. And there's always going to be 1 person who complains he wasn't told about it and it messes up his work.
5) Do it:
- Rename the organization droolsjbpm to kiegroup
- Make sure that no one can then create the organization droolsjbpm to break our old redirects. Contact github support for this?
-OLD WAY
- You can do one repository at a time, or do multiple at the same time.
-- I'd recommend to start with just 1 to get a feel for it.
- Do them in reverse order of the repository-list.txt!
- DELETE THE OLD REPOSITORY. DO NOT COPY IT, BUT MOVE IT.
-- After the move, we want anyone that's trying to pull from droolsjbpm to clearly *fail* pulling, but without losing local changes.
-- Some people will not have read your mail, they need to fail when pulling, otherwise they'll find themselves asking in 5 months why we haven't done any changes in 5 months.
- In droolsjbpm, create a repository called all_repositories_have_moved_to_github_com_kiegroup and make sure it's on the top of the list of the droolsjbpm repositories. Add a link in it's readme.adoc. This is for people who follow dead links from stackoverflow etc.
- Migrations like this should really happen during the weekend to reduce the impact on the rest of the team, ask your manager to swap a weekday for a weekend day.-
6) Clean up the fallout
- Go through your inventory and replace the references as far as humanly reasonable (stackoverflow is likely to be unreasonably, but openhub is not).
Anyway, that's my 2 cents (from my experience when moving svn to Git and splitting up the monastical build 5 years ago). Do this right and this will be almost a non-event (5 years ago it went pretty well on the engineering side, but less on the product side). Mess this up and this can be a disaster that will antagonize a lot of people...
was:
[~mark.proctor] asked me to make an inventory of what would need to happen if we ever migrate from github.com/droolsjbpm/ to github.com/kiegroup/. Technically this issue is very simply (just change it in GitHub), but organizationally this issue is difficult and dangerous: it needs to be done with care.
Preparation:
1) Get some experience
- Make a dummy repository in droolsjbpm, add some files and a few branches, fork it to your user account, add a PR on it, clone it 2 times:
-- Once from droolsjbpm
-- Once from your user account with droolsjbpm addes as upstream ("git remote add upstream ...github.com/droolsjbpm/...")
- Add some local changes on both dirs.
- On GitHub, go into the repository settings, in the danger zone and move it from droolsjbpm to kiegroup.
- Try using it from your 2 local clones ("git pull --rebase" or "git fetch upstream"). Copy paste the error messages for later use.
-- Fix it locally without recloning, without losing the local changes, etc. Record these steps clearly in a recipe to share with everyone else later.
-- Once it's fixed, see if you can do git operations without having to specify origin for the direct droolsjbpm clone. See also the git option "-u" - you'll likely have to add in the recipe as an extra step.
-- What happened to the open PR?
2) Clean up our repositories in droolsjbpm. There are a few dead ones that we should not be moving.
3) Set up roles and users of github.com/kiegroup.
4) Reach out
- Talk to each type of person involved:
-- R&D developer
-- QA
-- Productization
-- Product docs
-- Community
- Make an inventory of references to "github.com/droolsjbpm"
-- Do a find in path on all our files for "github.com/droolsjbpm")
-- Which webapps reference it?
--- jenkins
--- jira
--- www.openhub.net
--- gitter
--- stackoverflow
--- google forums / mailing lists
--- ...
- Set a date when you'll be moving the first repository.
-- Clearly communicate on all channels to let everyone know. This includes:
--- Public mailing lists and fora: Drools, OptaPlanner, jBPM, ...
--- Internal mailing lists: sme-brms, pm lists, bsig, etc
--- IRC channel topics
--- Social media (Twitter accounts etc)
--- ...
- Clearly communicate:
-- What you'll do
-- When you'll do it
-- How strong the freeze is: Can they make local changes meanwhile? Will they lose the open PR's? How will this affect their forks and branches? Answer all these questions proactively, based on your experience. Cropped screenshots are good!
-- A recipe: What they need to do - step by step for dummies - after the merge is done.
-- Get someone to review this mail before it goes out.
- You can't overcommunicate this, but you can undercommunicate it. And there's always going to be 1 person who complains he wasn't told about it and it messes up his work.
5) Do it:
- You can do one repository at a time, or do multiple at the same time.
-- I'd recommend to start with just 1 to get a feel for it.
- Do them in reverse order of the repository-list.txt!
- DELETE THE OLD REPOSITORY. DO NOT COPY IT, BUT MOVE IT.
-- After the move, we want anyone that's trying to pull from droolsjbpm to clearly *fail* pulling, but without losing local changes.
-- Some people will not have read your mail, they need to fail when pulling, otherwise they'll find themselves asking in 5 months why we haven't done any changes in 5 months.
- In droolsjbpm, create a repository called all_repositories_have_moved_to_github_com_kiegroup and make sure it's on the top of the list of the droolsjbpm repositories. Add a link in it's readme.adoc. This is for people who follow dead links from stackoverflow etc.
- Migrations like this should really happen during the weekend to reduce the impact on the rest of the team, ask your manager to swap a weekday for a weekend day.
6) Clean up the fallout
- Go through your inventory and replace the references as far as humanly reasonable (stackoverflow is likely to be unreasonably, but openhub is not).
Anyway, that's my 2 cents (from my experience when moving svn to Git and splitting up the monastical build 5 years ago). Do this right and this will be almost a non-event (5 years ago it went pretty well on the engineering side, but less on the product side). Mess this up and this can be a disaster that will antagonize a lot of people...
> Change github url from github.com/droolsjbpm/ to github.com/kiegroup/
> ---------------------------------------------------------------------
>
> Key: DROOLS-1048
> URL: https://issues.jboss.org/browse/DROOLS-1048
> Project: Drools
> Issue Type: Task
> Components: build
> Reporter: Geoffrey De Smet
> Assignee: Michael Biarnes Kiefer
> Priority: Critical
>
> [~mark.proctor] asked me to make an inventory of what would need to happen if we ever migrate from github.com/droolsjbpm/ to github.com/kiegroup/. Technically this issue is very simply (just change it in GitHub), but organizationally this issue is difficult and dangerous: it needs to be done with care.
> Preparation:
> 1) Get some experience
> - Make a dummy repository in droolsjbpm, add some files and a few branches, fork it to your user account, add a PR on it, clone it 2 times:
> -- Once from droolsjbpm
> -- Once from your user account with droolsjbpm addes as upstream ("git remote add upstream ...github.com/droolsjbpm/...")
> - Add some local changes on both dirs.
> - On GitHub, go into the repository settings, in the danger zone and move it from droolsjbpm to kiegroup.
> - Try using it from your 2 local clones ("git pull --rebase" or "git fetch upstream"). Copy paste the error messages for later use.
> -- Fix it locally without recloning, without losing the local changes, etc. Record these steps clearly in a recipe to share with everyone else later.
> -- Once it's fixed, see if you can do git operations without having to specify origin for the direct droolsjbpm clone. See also the git option "-u" - you'll likely have to add in the recipe as an extra step.
> -- What happened to the open PR?
> 2) Clean up our repositories in droolsjbpm. There are a few dead ones that we should not be moving.
> 3) Set up roles and users of github.com/kiegroup.
> 4) Reach out
> - Talk to each type of person involved:
> -- R&D developer
> -- QA
> -- Productization
> -- Product docs
> -- Community
> - Make an inventory of references to "github.com/droolsjbpm"
> -- Do a find in path on all our files for "github.com/droolsjbpm")
> -- Which webapps reference it?
> --- jenkins
> --- jira
> --- www.openhub.net
> --- gitter
> --- stackoverflow
> --- google forums / mailing lists
> --- ...
> - Set a date when you'll be moving the first repository.
> -- Clearly communicate on all channels to let everyone know. This includes:
> --- Public mailing lists and fora: Drools, OptaPlanner, jBPM, ...
> --- Internal mailing lists: sme-brms, pm lists, bsig, etc
> --- IRC channel topics
> --- Social media (Twitter accounts etc)
> --- ...
> - Clearly communicate:
> -- What you'll do
> -- When you'll do it
> -- How strong the freeze is: Can they make local changes meanwhile? Will they lose the open PR's? How will this affect their forks and branches? Answer all these questions proactively, based on your experience. Cropped screenshots are good!
> -- A recipe: What they need to do - step by step for dummies - after the merge is done.
> -- Get someone to review this mail before it goes out.
> - You can't overcommunicate this, but you can undercommunicate it. And there's always going to be 1 person who complains he wasn't told about it and it messes up his work.
> 5) Do it:
> - Rename the organization droolsjbpm to kiegroup
> - Make sure that no one can then create the organization droolsjbpm to break our old redirects. Contact github support for this?
> -OLD WAY
> - You can do one repository at a time, or do multiple at the same time.
> -- I'd recommend to start with just 1 to get a feel for it.
> - Do them in reverse order of the repository-list.txt!
> - DELETE THE OLD REPOSITORY. DO NOT COPY IT, BUT MOVE IT.
> -- After the move, we want anyone that's trying to pull from droolsjbpm to clearly *fail* pulling, but without losing local changes.
> -- Some people will not have read your mail, they need to fail when pulling, otherwise they'll find themselves asking in 5 months why we haven't done any changes in 5 months.
> - In droolsjbpm, create a repository called all_repositories_have_moved_to_github_com_kiegroup and make sure it's on the top of the list of the droolsjbpm repositories. Add a link in it's readme.adoc. This is for people who follow dead links from stackoverflow etc.
> - Migrations like this should really happen during the weekend to reduce the impact on the rest of the team, ask your manager to swap a weekday for a weekend day.-
> 6) Clean up the fallout
> - Go through your inventory and replace the references as far as humanly reasonable (stackoverflow is likely to be unreasonably, but openhub is not).
> Anyway, that's my 2 cents (from my experience when moving svn to Git and splitting up the monastical build 5 years ago). Do this right and this will be almost a non-event (5 years ago it went pretty well on the engineering side, but less on the product side). Mess this up and this can be a disaster that will antagonize a lot of people...
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 2 months