Codebase indexing not working since yesterday + weird thing about indexing

Hello, out of nowhere Codebase indexing stopped working with “Handshake failed”. I’m trying to reindex every 5 minutes since yesterday, in total I tried 100+ times + I reinstalled Cursor 3 times already.

Here is full log:

2024-10-08 09:58:45.287 [info] Indexing job successfully done or aborted.

2024-10-08 10:03:58.803 [info] Setting indexing intent to should-index

2024-10-08 10:03:58.812 [info] Creating merkle client.

2024-10-08 10:03:58.812 [info] Done creating merkle client.

2024-10-08 10:03:58.813 [info] Doing a startup handshake.

2024-10-08 10:03:58.843 [info] Finished initializing merkle tree in 32.21340000256896 ms.

2024-10-08 10:03:58.844 [info] Doing the initial handshake with hash: 19c67b49427b9e0807f082132dc15a6ced82ad3155ae5c75a3645ea5c16f267b

2024-10-08 10:03:58.845 [info] Handshake start

2024-10-08 10:03:59.421 [info] Handshake timing: 576.3827999979258

2024-10-08 10:04:01.433 [info] Handshake start

2024-10-08 10:04:01.584 [info] Handshake timing: 153.33520000055432

2024-10-08 10:04:05.596 [info] Handshake start

2024-10-08 10:04:05.759 [info] Handshake timing: 164.4362999983132

2024-10-08 10:04:13.771 [info] Handshake start

2024-10-08 10:04:14.002 [info] Handshake timing: 234.4172000028193

2024-10-08 10:04:30.011 [info] Handshake start

2024-10-08 10:04:30.189 [info] Handshake timing: 181.76240000128746

2024-10-08 10:05:02.196 [info] Handshake start

2024-10-08 10:05:02.353 [info] Handshake timing: 160.08930000290275

2024-10-08 10:06:06.361 [error] Handshake failed:

2024-10-08 10:06:06.363 [error] Error: timeout in handshake with retry

at t.RepoClientMultiplexer.handshakeWithRetry (c:\Users\admin\AppData\Local\Programs\cursor\resources\app\extensions\cursor-retrieval\dist\main.js:2:826553)

at async w.getServerStatus (c:\Users\admin\AppData\Local\Programs\cursor\resources\app\extensions\cursor-retrieval\dist\main.js:2:841881)

at async w.startIndexingRepository (c:\Users\admin\AppData\Local\Programs\cursor\resources\app\extensions\cursor-retrieval\dist\main.js:2:844181)

2024-10-08 10:06:06.363 [error] Handshake failed.

Version: 0.41.3
VSCode Version: 1.91.1
Commit: 51c8aff7cb5a89f4a0e462fbacab938bdbfaf140
Date: 2024-09-25T17:13:50.377Z
Electron: 29.4.0
ElectronBuildId: undefined
Chromium: 122.0.6261.156
Node.js: 20.9.0
V8: 12.2.281.27-electron.0
OS: Windows_NT x64 10.0.22631

WEIRD THING ON INDEXING:

Even when Indexing worked, I noticed that after you reindex the AI cannot find new files and thinks it’s missing from project root. Sometimes to the -point it tries to reinstall Next.js while it’s inside project root for weeks and project is like half done :smiley: I use re-index when I create a lot of files then ask composer to rescan /folder with @codebase flags as well. And in this way it misses couple of folders and files so I specify AI with full paths that these files are there, etc.

5 Likes

same here:
2024-10-08 15:35:48.470 [info] Indexing job successfully done or aborted.

2024-10-08 15:37:19.723 [info] Setting indexing intent to should-index

2024-10-08 15:37:19.723 [info] Creating merkle client.

2024-10-08 15:37:19.723 [info] Done creating merkle client.

2024-10-08 15:37:19.723 [info] Doing a startup handshake.

2024-10-08 15:37:19.825 [info] Finished initializing merkle tree in 101.54941700026393 ms.

2024-10-08 15:37:19.825 [info] Doing the initial handshake with hash: 99cddaff739cba664abddd23b3bec3c89751018895c5d28c40329780f2af22ac

2024-10-08 15:37:19.826 [info] Handshake start

2024-10-08 15:37:23.835 [info] Handshake start

2024-10-08 15:37:31.841 [info] Handshake start

2024-10-08 15:37:46.773 [info] Handshake start

2024-10-08 15:38:08.431 [info] Handshake start

2024-10-08 15:38:42.961 [info] Handshake start

2024-10-08 15:39:42.216 [info] Repo client: syncing embeddings with backend: https://repo42.cursor.sh

2024-10-08 15:39:53.533 [error] Handshake failed:

2024-10-08 15:39:53.533 [error] Error: timeout in handshake with retry

at t.RepoClientMultiplexer.handshakeWithRetry (/Applications/Cursor.app/Contents/Resources/app/extensions/cursor-retrieval/dist/main.js:2:826553)

at async w.getServerStatus (/Applications/Cursor.app/Contents/Resources/app/extensions/cursor-retrieval/dist/main.js:2:841881)

at async w.startIndexingRepository (/Applications/Cursor.app/Contents/Resources/app/extensions/cursor-retrieval/dist/main.js:2:844181)

2024-10-08 15:39:53.533 [error] Handshake failed.

2024-10-08 15:39:53.534 [info] Indexing job successfully done or aborted.

Version: 0.41.3
VSCode Version: 1.91.1
Commit: 51c8aff7cb5a89f4a0e462fbacab938bdbfaf140
Date: 2024-09-25T17:13:50.377Z
Electron: 29.4.0
ElectronBuildId: undefined
Chromium: 122.0.6261.156
Node.js: 20.9.0
V8: 12.2.281.27-electron.0
OS: Darwin arm64 24.0.0

3 Likes

I am on Ubuntu 22.04.

I have noticed that when you include Folder into composer, GPT still saw the files (or file paths) that were already deleted. So this might be a bug too!

So i wanted to “reindex” to see if that helps but, reindex didn’t help. Then i deleted the index and tried to create a new one.

And all i get is this famous “Handshake failed” error message too! Any solutions??

1 Like

Found logs as well now, if anyone wonders where are these logs, navigate to In “bottom panel” / “Output” tab and then select “Cursor indexing & retrieval” on the right side. Here is my log output:

2024-10-08 10:50:43.947 [info] Setting indexing intent to should-index

2024-10-08 10:50:43.948 [info] Creating merkle client.

2024-10-08 10:50:43.948 [info] Done creating merkle client.

2024-10-08 10:50:43.948 [info] Doing a startup handshake.

2024-10-08 10:50:43.951 [info] Finished initializing merkle tree in 3.104768991470337 ms.

2024-10-08 10:50:43.951 [info] Doing the initial handshake with hash: my_hash …

2024-10-08 10:50:43.952 [info] Handshake start

2024-10-08 10:50:44.071 [info] Handshake timing: 119.70853996276855

2024-10-08 10:50:46.071 [info] Handshake start

2024-10-08 10:50:46.191 [info] Handshake timing: 119.40664315223694

2024-10-08 10:50:50.191 [info] Handshake start

2024-10-08 10:50:50.311 [info] Handshake timing: 120.18391513824463

2024-10-08 10:50:58.311 [info] Handshake start

2024-10-08 10:50:58.431 [info] Handshake timing: 119.539146900177

2024-10-08 10:51:14.433 [info] Handshake start

2024-10-08 10:51:14.551 [info] Handshake timing: 120.00630402565002

2024-10-08 10:51:46.553 [info] Handshake start

2024-10-08 10:51:46.671 [info] Handshake timing: 119.28063488006592

2024-10-08 10:52:50.672 [error] Handshake failed:

2024-10-08 10:52:50.672 [error] Error: timeout in handshake with retry

at t.RepoClientMultiplexer.handshakeWithRetry (/tmp/.mount_cursorSz3Vdf/resources/app/extensions/cursor-retrieval/dist/main.js:2:826553)

at async w.getServerStatus (/tmp/.mount_cursorSz3Vdf/resources/app/extensions/cursor-retrieval/dist/main.js:2:841881)

at async w.startIndexingRepository (/tmp/.mount_cursorSz3Vdf/resources/app/extensions/cursor-retrieval/dist/main.js:2:844181)

2024-10-08 10:52:50.672 [error] Handshake failed.

2024-10-08 10:52:50.672 [info] Indexing job successfully done or aborted.

4 Likes

Having the exact same

2024-10-08 05:51:27.427 [error] Handshake failed:
2024-10-08 05:51:27.429 [error] Error: timeout in handshake with retry
    at t.RepoClientMultiplexer.handshakeWithRetry (/Applications/Cursor.app/Contents/Resources/app/extensions/cursor-retrieval/dist/main.js:2:826553)
    at w.getServerStatus (/Applications/Cursor.app/Contents/Resources/app/extensions/cursor-retrieval/dist/main.js:2:841881)
    at w.startIndexingRepository (/Applications/Cursor.app/Contents/Resources/app/extensions/cursor-retrieval/dist/main.js:2:844181)
2024-10-08 05:51:27.429 [error] Handshake failed.
2024-10-08 05:51:27.429 [info] Indexing job successfully done or aborted.

issue

Now when i use ctrl + l for chat i get additional errors:
Error indexing codebase
Codebase chat is falling back to BM25, which is slower and less accurate than embeddings.

Same issue here. Any solutions? This is a blocker, which makes product unusable.

5 Likes

Same issue !!!

4 Likes

same issue!! Restart cursor not working!

Same issue here.

Same issue, call for emergency fix

1 Like

same issue here.

same issue here if indexing not working this consider as normal ai tools

Welcome to the party!
I hope this time this f**** issue will be addressed and solved definitively

Raises $60M, quality and stability drops quite dramatically. Pretty wild.

At least they know where to spend their money

Fix it please, I’m losing working hours, holy cow!!!

2 Likes

I got this error on my Macbook yesterday and thought it was something about the settings on that device, however, my PC is getting it now too.

Same issue on my end as well. Pretty small codebase too.

A statement on this would also be appreciated.

That would be so amazing. Confirm there is an issue and explain what not having working indexing impacts.