Bug 515380

Summary: SMB workers killed during large file transfers causing 40% stall
Product: [Frameworks and Libraries] kio-extras Reporter: Christofle <defran.christofle>
Component: SambaAssignee: Plasma Bugs List <plasma-bugs-null>
Status: RESOLVED DOWNSTREAM    
Severity: normal CC: sitter
Priority: NOR    
Version First Reported In: 25.04.3   
Target Milestone: ---   
Platform: Debian stable   
OS: Linux   
Latest Commit: Version Fixed/Implemented In:
Sentry Crash Report:
Attachments: Filtered logs

Description Christofle 2026-02-01 11:19:07 UTC
Created attachment 189130 [details]
Filtered logs

=== SYSTEM INFORMATION ===
OS: Debian GNU/Linux 13 (trixie)
Kernel: 6.12.63+deb13-amd64
Desktop: KDE Plasma (X11 and Wayland tested)
Dolphin: 25.04.3
kio-extras: 4:25.04.3-1
libsmbclient: 2:4.22.6+dfsg-0+deb13u1
Network: Gigabit Ethernet (Intel igc driver)

=== PROBLEM DESCRIPTION ===
SMB worker processes are killed during large file transfers (>500MB), causing transfers to fail around 40% completion.

=== REPRODUCTION STEPS ===
1. Open Dolphin
2. Navigate to Network → SMB → any SMB server (tested with multiple Synology NAS)
3. Attempt to copy a file >500MB from local disk to SMB share
4. Observe transfer behavior

=== OBSERVED BEHAVIOR ===
- Transfer starts at ~10 Mbit/s
- Speed drops to near 0 within seconds
- Progress stalls around 40%
- Multiple SMB workers spawn (6+) and are killed by KIO
- Log shows: "kf.kio.core: killing worker process pid XXXX"
- Followed by: "Socket not connected QLocalSocket::PeerClosedError"
- Then: "An error occurred during write. The worker terminates now."

=== EXPECTED BEHAVIOR ===
File should transfer completely at full network speed.

=== WORKAROUNDS THAT WORK ===
1. Direct smbclient: Transfers complete at 68 MB/s in ~40 seconds
2. CIFS mount in /etc/fstab: Transfers work perfectly at full speed
3. GVFS (Nautilus): Also fails with same symptoms (24 KB/s)

=== ADDITIONAL CONTEXT ===
- Problem appeared after accidental KDE uninstall/reinstall
- Tested on both X11 and Wayland sessions - identical behavior
- Problem occurs with multiple different SMB servers (Synology NAS)
- Network verified working (1 Gbps full duplex, no packet loss)
- Same user account works perfectly from Windows clients

=== CONCLUSION ===
This appears to be a worker management issue in KIO where the framework incorrectly terminates SMB workers during active transfers, possibly due to timeout miscalculation or resource limit errors.
Comment 1 Harald Sitter 2026-02-01 22:21:33 UTC
Thank you for your bug report!
Debian advises users not to submit bugs upstream (https://www.debian.org/Bugs/Reporting), as this version of the KDE software no longer receives updates or maintenance from KDE, and so it's possible that the bug exists only in Debian at this point.
Could you report the bug to Debian using the report bug utility (https://packages.debian.org/stable/utils/reportbug)? If necessary, the maintainer of the package will forward the bug upstream.
Thanks for understanding!