Git clone slowness after upgrading to Elite plan

[nebil]

Hi,
we have recently requested the Elite plan trial.

We have noticed a significant improvement on the build, however we also noticed sometime the git clone steps takes ages to run (more then 25 minutes). This used to always take no more then 90 seconds to compete, but after we joined the Elite plan we see at least 1/5 builds taking an excessive amount of time. Here’s an example:

Bitrise - Mobile Continuous Integration and Delivery

Our repo is over 1gb and we are not using any cache for this. However we can’t understand why this would take so long. Can you provide more clarity on this?

Thanks

[godreikrisztian]

Hi @nebil,
could you please also share an example build log, where the git clone is significantly faster?

[nebil]

Hi @godreikrisztian thanks for your message.
here’s the log part from this morning:

±-----------------------------------------------------------------------------+
| (1) git-clone |
±-----------------------------------------------------------------------------+
| id: git-clone |
| version: 4.0.11 |
| collection: https://github.com/bitrise-io/bitrise-steplib.git |
| toolkit: go |
| time: 2018-05-14T06:44:46-07:00 |
±-----------------------------------------------------------------------------+
| |
e[36mINFOe[0m[06:44:46] * e[32;1m[OK]e[0m Step dependency (go) installed, available.
e[34;1mconfig:
e[0m

• RepositoryURL: git@bitbucket.org:depop/depop-ios.git
• CloneIntoDir: /Users/vagrant/git
• Commit: ee9f90c2a6f3
• Tag:
• Branch: feature/DD-5942
• BranchDest: dev
• PRID: 2126
• PRRepositoryURL: git@bitbucket.org:depop/depop-ios.git
• PRMergeBranch:
• ResetRepository: false
• CloneDepth: 0
• BuildURL: XXXX
• BuildAPIToken: XXXXXX
• UpdateSubmodules: true
• ManualMerge: true
  e[34;1mgit "init"e[0m
  Initialized empty Git repository in /Users/vagrant/git/.git/
  e[34;1mgit “remote” “add” “origin” "git@bitbucket.org:depop/depop-ios.git"e[0m
  e[34;1mgit “fetch” “origin” "dev"e[0m
  Warning: Permanently added ‘bitbucket.org,104.192.143.2’ (RSA) to the list of known hosts.
  From bitbucket.org:depop/depop-ios

• branch dev -> FETCH_HEAD
• [new branch] dev -> origin/dev
  e[34;1mgit “checkout” "dev"e[0m
  Switched to a new branch ‘dev’
  Branch ‘dev’ set up to track remote branch ‘dev’ from ‘origin’.
  e[34;1mgit “merge” "origin/dev"e[0m
  Already up to date.
  commit hash: 3e8809ef0c99f0e8d6551aca8a530d2a88478f93
  e[34;1mgit “fetch” “origin” "feature/DD-5942"e[0m
  Warning: Permanently added ‘bitbucket.org,104.192.143.2’ (RSA) to the list of known hosts.
  From bitbucket.org:depop/depop-ios
• branch feature/DD-5942 -> FETCH_HEAD
• [new branch] feature/DD-5942 -> origin/feature/DD-5942
  e[34;1mgit “merge” "ee9f90c2a6f3"e[0m
  Updating 3e8809ef0…ee9f90c2a
  Fast-forward
  [these lines were removed]
  4 files changed, 4 insertions(+), 10 deletions(-)
  e[34;1mgit “submodule” “update” “–init” "–recursive"e[0m
  e[34;1mgit “checkout” "–detach"e[0m
  HEAD is now at ee9f90c2a… do not translate ‘Terms of Service’ and ‘Privacy Policy’ in the app
  e[34;1m
  Exporting git logs
  e[0m
  => GIT_CLONE_COMMIT_MESSAGE_SUBJECT
  value: XXXXXXXX

=> GIT_CLONE_COMMIT_MESSAGE_BODY
value:

=> GIT_CLONE_COMMIT_AUTHOR_NAME
value: XXXXX

=> GIT_CLONE_COMMIT_AUTHOR_EMAIL
value: XXXX@depop.com

=> GIT_CLONE_COMMIT_COMMITER_NAME
value: XXXXXX

=> GIT_CLONE_COMMIT_COMMITER_EMAIL
value: XXXX@depop.com

=> GIT_CLONE_COMMIT_HASH
value: ee9f90c2a6f33bab7f8f34589fb6499cf078f700

=> GIT_CLONE_COMMIT_COUNT
value: 18611

e[32;1m
Successe[0m
| |
±–±--------------------------------------------------------------±---------+
| e[32;1m✓e[0m | e[32;1mgit-clonee[0m | 108 sec |
±–±--------------------------------------------------------------±---------+

[godreikrisztian]

Hi @nebil,

this is a significant speed difference between the two builds (27 min vs 108 sec).

Both build happened on the Elite machines? Is it possible that it was just a temporary issue on Bitbucket, do you regularly see this slow git clones?

[nebil]

hi @godreikrisztian yes this is big issue we have only experienced once we migrated to the Elite plan. is it possible that those machines have a different access to the network, more subjected to random latency?

thank you

[godreikrisztian]

@nebil could you please write us a message on bitrise.io on-site-chat? We need more information to identify the issue.