Skip to content
代码片段 群组 项目
未验证 提交 c83792f2 编辑于 作者: Brendan Lynch's avatar Brendan Lynch 提交者： GitLab
浏览文件

Moved import project troubleshooting to separate page 


Co-authored-by: default avatarSuzanne Selhorn <sselhorn@gitlab.com>
上级 3dc75b1b
No related branches found
No related tags found
无相关合并请求
隐藏空白变更内容
行内 左右并排
doc/user/project/import/github.md
+ 2
− 108
浏览文件 @ c83792f2
......@@ -183,14 +183,14 @@ When the **Organization** tab is selected, you can further narrow down your sear
To make imports as fast as possible, the following items aren't imported from GitHub by default:
- More than approximately 30,000 comments because of a [limitation of the GitHub API](#missing-comments).
- More than approximately 30,000 comments because of a [limitation of the GitHub API](troubleshooting_github_import.md#missing-comments).
- Markdown attachments from repository comments, release posts, issue descriptions, and pull request descriptions. These can include
images, text, or binary attachments. If not imported, links in Markdown to attachments break after you remove the attachments from GitHub.
You can choose to import these items, but this could significantly increase import time. To import these items, select the appropriate fields in the UI:
- **Use alternative comments import method**. If importing GitHub projects with more than approximately 30,000 comments across all issues and pull requests, you should enable this method because of a
[limitation of the GitHub API](#missing-comments).
[limitation of the GitHub API](troubleshooting_github_import.md#missing-comments).
- **Import Markdown attachments**.
- **Import collaborators** (selected by default). Leaving it selected might result in new users using a seat in the group or namespace,
and being granted permissions [as high as project owner](#collaborators-members). Only direct collaborators are imported.
......@@ -485,109 +485,3 @@ LoadModule ssl_module lib/httpd/modules/mod_ssl.so
Header edit* Link "https://GITHUB_ENTERPRISE_HOSTNAME" "https://PROXY_HOSTNAME"
</VirtualHost>
```
## Troubleshooting
### Manually continue a previously failed import process
In some cases, the GitHub import process can fail to import the repository. This causes GitLab to abort the project import process and requires the
repository to be imported manually. Administrators can manually import the repository for a failed import process:
1. Open a Rails console.
1. Run the following series of commands in the console:
```ruby
project_id = <PROJECT_ID>
github_access_token = <GITHUB_ACCESS_TOKEN>
github_repository_path = '<GROUP>/<REPOSITORY>'
github_repository_url = "https://#{github_access_token}@github.com/#{github_repository_path}.git"
# Find project by ID
project = Project.find(project_id)
# Set import URL and credentials
project.import_url = github_repository_url
project.import_type = 'github'
project.import_source = github_repository_path
project.save!
# Create an import state if the project was created manually and not from a failed import
project.create_import_state if project.import_state.blank?
# Set state to start
project.import_state.force_start
# Optional: If your import had certain optional stages selected or a timeout strategy
# set, you can reset them here. Below is an example.
# The params follow the format documented in the API:
# https://docs.gitlab.com/ee/api/import.html#import-repository-from-github
Gitlab::GithubImport::Settings
.new(project)
.write(
timeout_strategy: "optimistic",
optional_stages: {
single_endpoint_issue_events_import: true,
single_endpoint_notes_import: true,
attachments_import: true,
collaborators_import: true
}
)
# Trigger import from second step
Gitlab::GithubImport::Stage::ImportRepositoryWorker.perform_async(project.id)
```
### Import fails due to missing prefix
In GitLab 16.5 and later, you might get an error that states `Import failed due to a GitHub error: (HTTP 406)`.
This issue occurs because, in GitLab 16.5, the path prefix `api/v3` was removed from the GitHub importer. This happened because the importer stopped using the `Gitlab::LegacyGithubImport::Client`. This client automatically added the `api/v3` prefix on imports from a GitHub Enterprise URL.
To work around this error, [add the `api/v3` prefix](https://gitlab.com/gitlab-org/gitlab/-/issues/438358#note_1978902725) when importing from a GitHub Enterprise URL.
### Errors when importing large projects
The GitHub importer might encounter some errors when importing large projects.
#### Missing comments
The GitHub API has a limit that prevents more than approximately 30,000 notes or diff notes from being imported.
When this limit is reached, the GitHub API instead returns the following error:
```plaintext
In order to keep the API fast for everyone, pagination is limited for this resource. Check the rel=last link relation in the Link response header to see how far back you can traverse.
```
If you are importing GitHub projects with a large number of comments, you should select the **Use alternative comments import method**
[additional item to import](#select-additional-items-to-import) checkbox. This setting makes the import process take longer because it increases the number of network requests
required to perform the import.
#### Reduce GitHub API request objects per page
Some GitHub API endpoints might return a `500` or `502` error for project imports from large repositories.
To reduce the chance of these errors, in the group project importing the data, enable the
`github_importer_lower_per_page_limit` feature flag. When enabled, the flag reduces the
page size from `100` to `50`.
To enable this feature flag:
1. Start a [Rails console](../../../administration/operations/rails_console.md#starting-a-rails-console-session).
1. Run the following `enable` command:
```ruby
group = Group.find_by_full_path('my/group/fullpath')
# Enable
Feature.enable(:github_importer_lower_per_page_limit, group)
```
To disable the feature flag, run this command:
```ruby
# Disable
Feature.disable(:github_importer_lower_per_page_limit, group)
```
### GitLab instance cannot connect to GitHub
Self-managed instances that run GitLab 15.10 or earlier, and are behind proxies, cannot resolve DNS for `github.com` or `api.github.com`.
In this situation, the GitLab instance fails to connect to GitHub during the import and you must add `github.com` and `api.github.com`
entries in the [allowlist for local requests](../../../security/webhooks.md#allow-outbound-requests-to-certain-ip-addresses-and-domains).
doc/user/project/import/troubleshooting_github_import.md 0 → 100644
+ 113
− 0
浏览文件 @ c83792f2
---
stage: Manage
group: Import and Integrate
info: To determine the technical writer assigned to the Stage/Group associated with this page, see https://handbook.gitlab.com/handbook/product/ux/technical-writing/#assignments
---
## Troubleshooting importing a project from GitHub to GitLab
When importing a project from GitHub to GitLab, you might encounter the following issues.
### Manually continue a previously failed import process
In some cases, the GitHub import process can fail to import the repository. This causes GitLab to abort the project import process and requires the
repository to be imported manually. Administrators can manually import the repository for a failed import process:
1. Open a Rails console.
1. Run the following series of commands in the console:
```ruby
project_id = <PROJECT_ID>
github_access_token = <GITHUB_ACCESS_TOKEN>
github_repository_path = '<GROUP>/<REPOSITORY>'
github_repository_url = "https://#{github_access_token}@github.com/#{github_repository_path}.git"
# Find project by ID
project = Project.find(project_id)
# Set import URL and credentials
project.import_url = github_repository_url
project.import_type = 'github'
project.import_source = github_repository_path
project.save!
# Create an import state if the project was created manually and not from a failed import
project.create_import_state if project.import_state.blank?
# Set state to start
project.import_state.force_start
# Optional: If your import had certain optional stages selected or a timeout strategy
# set, you can reset them here. Below is an example.
# The params follow the format documented in the API:
# https://docs.gitlab.com/ee/api/import.html#import-repository-from-github
Gitlab::GithubImport::Settings
.new(project)
.write(
timeout_strategy: "optimistic",
optional_stages: {
single_endpoint_issue_events_import: true,
single_endpoint_notes_import: true,
attachments_import: true,
collaborators_import: true
}
)
# Trigger import from second step
Gitlab::GithubImport::Stage::ImportRepositoryWorker.perform_async(project.id)
```
### Import fails due to missing prefix
In GitLab 16.5 and later, you might get an error that states `Import failed due to a GitHub error: (HTTP 406)`.
This issue occurs because, in GitLab 16.5, the path prefix `api/v3` was removed from the GitHub importer. This happened because the importer stopped using the `Gitlab::LegacyGithubImport::Client`. This client automatically added the `api/v3` prefix on imports from a GitHub Enterprise URL.
To work around this error, [add the `api/v3` prefix](https://gitlab.com/gitlab-org/gitlab/-/issues/438358#note_1978902725) when importing from a GitHub Enterprise URL.
### Errors when importing large projects
The GitHub importer might encounter some errors when importing large projects.
#### Missing comments
The GitHub API has a limit that prevents more than approximately 30,000 notes or diff notes from being imported.
When this limit is reached, the GitHub API instead returns the following error:
```plaintext
In order to keep the API fast for everyone, pagination is limited for this resource. Check the rel=last link relation in the Link response header to see how far back you can traverse.
```
When importing GitHub projects with a large number of comments, select the **Use alternative comments import method**
[additional item to import](github.md#select-additional-items-to-import) checkbox. This setting makes the import process take longer because it increases the number of network requests
required to perform the import.
#### Reduce GitHub API request objects per page
Some GitHub API endpoints might return a `500` or `502` error for project imports from large repositories.
To reduce the chance of these errors, in the group project importing the data, enable the
`github_importer_lower_per_page_limit` feature flag. When enabled, the flag reduces the
page size from `100` to `50`.
To enable this feature flag:
1. Start a [Rails console](../../../administration/operations/rails_console.md#starting-a-rails-console-session).
1. Run the following `enable` command:
```ruby
group = Group.find_by_full_path('my/group/fullpath')
# Enable
Feature.enable(:github_importer_lower_per_page_limit, group)
```
To disable the feature flag, run this command:
```ruby
# Disable
Feature.disable(:github_importer_lower_per_page_limit, group)
```
### GitLab instance cannot connect to GitHub
Self-managed instances that run GitLab 15.10 or earlier, and are behind proxies, cannot resolve DNS for `github.com` or `api.github.com`.
The GitLab instance fails to connect to GitHub during the import and you must add `github.com` and `api.github.com`
entries in the [allowlist for local requests](../../../security/webhooks.md#allow-outbound-requests-to-certain-ip-addresses-and-domains).
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或