Linda Walsh
2012-Aug-05 08:50 UTC
rsync 3.0.9 incompatible with self? (proto incompat on local->local)
Familiar? protocol error on a local file copy? How can I have a proto incompat when it it is talking to itself?... This was on the ERROUT... rsync: writefd_unbuffered failed to write 4 bytes to socket [sender]: Broken pipe (32) File-list index 531530 not in 531937 - 533437 (finish_hard_link) [generator] rsync error: protocol incompatibility (code 2) at rsync.c(652) [generator=3.0.9] rsync: connection unexpectedly closed (54286 bytes received so far) [sender] rsync error: error in rsync protocol data stream (code 12) at io.c(605) [sender=3.0.9] args used (approx, as they were passed in perl)... --archive --8-bit-output --human-readable --acls --xattrs --hard-links --whole-file --one-file-system --compare-dest=/home/. --exclude={**/.recycle/**,bhome/la**,bhome/athen/**, bhome/root**,bhome/mar** CPAN-ishtar-build-cache/**} /home/snapdir/old_snapdir/. /home.diff/
Linda Walsh
2012-Aug-07 20:56 UTC
rsync 3.0.9 incompatible with self? (proto incompat on local->local)
This list seems awfully quiet...
Linda Walsh
2012-Aug-07 23:08 UTC
rsync 3.0.9 incompatible with self? (source of problem found tnx to TimeMachinethread)...
This bug is caused by the incremental recursion added in 3.0. That's the incompatibility -- incremental recursion is incompatible with keeping the Hard links... ... I didn't know hard links were so difficult for rsync...now I know why... they don't pre-read the tree anymore...so there's no way they can really keep track of what files need to be hard-linked -- thus the hardlink generator fails because incremental-recursion is incompatible with doing this... How LAME!... if that's the case, then -H should auto-disable incremental recursion.... (which is what I did... I reran rsync with a different dataset and got another failure like below, then I tried it with another dataset (daily diffs I can only keep so long)... and it ran to completion. Well...at least I know how to work around it...but until they get incremental recursion fixed, it should probably be auto-disabled for -H usage. My storage requirements would seriously go up with hard links ignored. (especially since I run deduplication on my files now and then that links duplicates together!).... Linda Walsh wrote:> Familiar? protocol error on a local file copy? > How can I have a proto incompat when it it is talking to itself?... > > This was on the ERROUT... > > rsync: writefd_unbuffered failed to write 4 bytes to socket [sender]: > Broken pipe (32) > File-list index 531530 not in 531937 - 533437 (finish_hard_link) > [generator] > rsync error: protocol incompatibility (code 2) at rsync.c(652) > [generator=3.0.9] > rsync: connection unexpectedly closed (54286 bytes received so far) > [sender] > rsync error: error in rsync protocol data stream (code 12) at io.c(605) > [sender=3.0.9] > > > > args used (approx, as they were passed in perl)... > > --archive --8-bit-output --human-readable > --acls --xattrs --hard-links > --whole-file --one-file-system > --compare-dest=/home/. > --exclude={**/.recycle/**,bhome/la**,bhome/athen/**, > bhome/root**,bhome/mar** > CPAN-ishtar-build-cache/**} > /home/snapdir/old_snapdir/. /home.diff/
Wayne Davison
2012-Aug-12 17:44 UTC
rsync 3.0.9 incompatible with self? (proto incompat on local->local)
On Sun, Aug 5, 2012 at 1:50 AM, Linda Walsh <rsync at tlinx.org> wrote:> How can I have a proto incompat when it it is talking to itself?... >You cited a hard-link bug in the incremental recursion algorithm. You can work around it by specifying --no-inc-recursive (--no-i-r). If you have a reproducible test case, I'd like to see it (if possible) to do some debugging. ..wayne.. -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.samba.org/pipermail/rsync/attachments/20120812/4c9e3422/attachment.html>
Reasonably Related Threads
- smbd: PANIC (pid xxxxx): internal error -- ? causes?
- DO NOT REPLY [Bug 5407] New: hlink.c:480: finish_hard_link: Assertion `flist != ((void *)0)' failed.
- link(2) EMLINK error behavior with --link-dest and --hard-links
- why is my "nmbd" confused about network interfaces?
- DO NOT REPLY [Bug 7454] New: assertion failed in finish_hard_link()