-
Notifications
You must be signed in to change notification settings - Fork 340
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
NodeReleaser not triggered #434
Comments
is the FS user closing the files? Can you see RELEASE opcodes in the debug output? |
Yes, when debug is set to true for the mount-options, I see the RELEASE being triggered, but it never runs my function. |
2022/07/12 18:00:45 rx 78: READ n2 {Fh 0 [3538944 +131072) L 0 RDONLY,0x8000} |
the code is pretty straightforward, see Line 814 in d01fda7
are you sure the inode implementation ( |
Well, I created it the same way as the GETATTR and that one works fine. READDIR and LOOKUP also works fine. //WORKS //DOES NOT WORK |
if you have no file handler, there is nothing to release. You can hook on Forget instead, but if your internals are tied a HTTP body that is read-once, you should use file handles, or things will get weird with concurrent access to the same file. |
Do you have any example how to use OPEN and READ with a http request? I tried, but can't get it to work correctly with file handlers. Exactly as you described, by READs happens in random order :( |
I have this, and it's working 90% of the time. But 10% of the time, the md5sum of the resulting file does not match the original one. Obviously I am not understanding it correctly :) The Read function is not always being executed in the "correct" order even though I have set SingleThreaded: true
|
I don't understand why the READ is called in almost random order over the file even though it's like one single read. Is there a way to force it to READ in "order" or do I need to rely on the HTTP server to accept "range" requests? 2022/07/13 04:49:23 OFF 0 PART 0 |
The kernel has caching and read-ahead, so it sometimes tries to be clever and read ahead of what is requested. |
Yes, that's kernel read-ahead. Used to be a tricky problem (rfjakob/gocryptfs#92), but, now we have the Set it to true like here: |
That would be awesome! Unfortunately, that flag does not seem to exist? Was this added in some other branch? Also can't find it in the documentation for MountOptions:
|
Ah! I see now, this was added in late 2021 and was never included in any release. |
@rfjakob Thanks!! Now it's working the way it should! Even works with multi-threading enabled. |
My Release function is never triggered. Other events works fine, like GetAttr, Lookup etc.
I have debug enabled and see that FUSE triggers the Release event, but my function is never run
The text was updated successfully, but these errors were encountered: