samba-bugs at samba.org
2011-Dec-30 01:02 UTC
[Bug 8682] New: Skip current transfer keyboard function
https://bugzilla.samba.org/show_bug.cgi?id=8682 Summary: Skip current transfer keyboard function Product: rsync Version: 3.1.0 Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P5 Component: core AssignedTo: wayned at samba.org ReportedBy: b1566587 at nwldx.com QAContact: rsync-qa at samba.org rsync should stop transferring the active file upon press of a specific key. There is already extensive file exclusion functionality, but the ability to skip the current file (the one visible when using --progress) would be really quite handy to have. -- Configure bugmail: https://bugzilla.samba.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug.
samba-bugs at samba.org
2015-Sep-18 09:20 UTC
[Bug 8682] Skip current transfer keyboard function
https://bugzilla.samba.org/show_bug.cgi?id=8682 --- Comment #1 from martin f. krafft <bugzilla.samba.org at pobox.madduck.net> --- A hack is to save the file on the backedu up host, and overwrite the inode, e.g. "echo > file". This will then cause BackupPC to move on and log an error. But indeed, having the information about - which file is being processed - some metadata, e.g. size - when processing started available in the web interface would be great. Ideally, this would then be augmented with the ability to cancel/skip this file and optionally to add the file or the containing directory to the global or host exclude list. -- You are receiving this mail because: You are the QA Contact for the bug.
samba-bugs at samba.org
2015-Sep-18 09:20 UTC
[Bug 8682] Skip current transfer keyboard function
https://bugzilla.samba.org/show_bug.cgi?id=8682 --- Comment #2 from martin f. krafft <bugzilla.samba.org at pobox.madduck.net> --- (the hack is courtesy of Ronny Aasen) -- You are receiving this mail because: You are the QA Contact for the bug.
samba-bugs at samba.org
2015-Sep-19 10:42 UTC
[Bug 8682] Skip current transfer keyboard function
https://bugzilla.samba.org/show_bug.cgi?id=8682 --- Comment #3 from roland <devzero at web.de> --- that would be a quite useful feature. i`d recommend that this feature should be activated with a commandline option for safety purpose. and skipping should then be repeated while the key is pressed down, so one could skip multiple files at once or an entire directory structure (e.g. when accidentally syncing /proc etc) -- You are receiving this mail because: You are the QA Contact for the bug.
samba-bugs at samba.org
2018-Mar-22 01:54 UTC
[Bug 8682] Skip current transfer keyboard function
https://bugzilla.samba.org/show_bug.cgi?id=8682 --- Comment #4 from Christian Kujau <lists at nerdbynature.de> --- If this ever gets implemented: instead of (interactively) pressing a key to interrupt the current transfer of a particular object, I'd like it to also react to a signal (e.g. SIGUSR1) that can be sent to rsync, even when it's not started interactively. -- You are receiving this mail because: You are the QA Contact for the bug.
samba-bugs at samba.org
2022-Mar-04 10:34 UTC
[Bug 8682] Skip current transfer keyboard function
https://bugzilla.samba.org/show_bug.cgi?id=8682 --- Comment #6 from Joachim Wagner <jwagner at computing.dcu.ie> --- The "echo > sourcefile" workaround seems to not trigger an error, at least not straight away and at least on XFS, but instead speeds up the operation and stops writing more data to the target file. Observation (both source and target filesystems are local, 5.7 TB file at 68% at time of intervention): * Before "echo > sourcefile", rsync -P reported about 200MB/s * After "echo > sourcefile", rsync -P reports about 650MB/s and the target file's mtime and size no longer change * At 100%, rsync says "read errors mapping "redacted": No data available (61) WARNING: redacted failed verification -- update discarded (will try again)." and then syncs the 1-byte file created with `echo`. Still grateful @martin for sharing the idea. I found this searching for a way to inject an I/O error. For the source file backup, `cp --reflink=always` is your friend. -- You are receiving this mail because: You are the QA Contact for the bug.