Bug 723370 - rsync - IO errors and Kernel crash
Summary: rsync - IO errors and Kernel crash
Keywords:
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: Fedora
Classification: Fedora
Component: rsync
Version: 15
Hardware: i686
OS: Linux
unspecified
unspecified
Target Milestone: ---
Assignee: Vojtech Vitek
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2011-07-20 01:29 UTC by nomnex
Modified: 2015-03-04 23:57 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2011-08-22 10:25:33 UTC
Type: ---


Attachments (Terms of Use)
phone photo of the message after kernel crash (159.82 KB, image/jpeg)
2011-07-31 04:46 UTC, nomnex
no flags Details

Description nomnex 2011-07-20 01:29:41 UTC
Description of problem:

Vanilla F-15 lxde and grsync. I have tried grsync for the first time on F-15, to backup some files on my PC hhd to my USB hdd.

The fist time: I generated a kernel crash:
Kernel BUG at fs/ext4/inode.c:2188!
Invalid opcode: 0000 [#1] SMP

The second time: I tried another backup (PC hdd > USB hdd) speed was *very* slow
avi. 1.08MB/s (first movie - OKAY)
avi. 558.02kB/s (second movie - Warning message)

Grsync generated a warning message midway in the second .avi backup, and did not complete the operation (msg):

rsync: writefd_unbuffered failed to write 4 bytes to socket [sender]: Broken pipe (32)
rsync: write failed on "/media/disk2/torrent/avi/0004.avi": Read-only file system (30)
rsync error: error in file IO (code 11) at receiver.c(322) [receiver=3.0.8]
rsync: connection unexpectedly closed (70 bytes received so far) [sender]
rsync error: error in rsync protocol data stream (code 12) at io.c(601) [sender=3.0.8]
Rsync process exit status: 12

Version-Release number of selected component (if applicable):

Grsync 1.1.1
2.6.38.8-35.fc15.i686

How reproducible:

Always

Steps to Reproduce:
1. Plug USB HD
2. Backup movie (1 GB) using RSYNC from local hdd to remote USB hdd
3. 
  
Actual results:

Kernel crash or Grsync warning message

Expected results:

To backup on my USB HD as it used to be.

Additional info:

Explain me what command to pass to send the kernel backtrace of the crash as an attachment. When the kernel crash occurred, the bsd? (black screen of the death) msg said: ABRT started. Is the output in var/log, and what file is it? Thanks in advance.

# /var/log
anaconda           cron-20110717      messages-20110710  spooler-20110710
audit              cups               messages-20110717  spooler-20110717
boot.log           dracut.log         ntpstats           sssd
boot.log-20110704  lastlog            ppp                tallylog
boot.log-20110710  lxdm.log           prelink            wpa_supplicant.log
boot.log-20110717  lxdm.log.old       secure             wtmp
btmp               maillog            secure-20110704    Xorg.0.log
btmp-20110701      maillog-20110704   secure-20110710    Xorg.0.log.old
ConsoleKit         maillog-20110710   secure-20110717    Xorg.9.log
cron               maillog-20110717   setroubleshoot     yum.log
cron-20110704      messages           spooler
cron-20110710      messages-20110704  spooler-20110704

Comment 1 nomnex 2011-07-20 01:34:54 UTC
Pingback forum post with a similar issue
http://www.fedoraforum.org/forum/showthread.php?p=1496705

Comment 2 nomnex 2011-07-20 02:10:50 UTC
I tested with a rsync command (with minimal option). It was fast and went fine.

[mt@nh28d ~]$ rsync -vvr /home/mt/download/KinDzaDza /media/disk2/torrent/movie/
sending incremental file list
delta-transmission disabled for local transfer or --whole-file
KinDzaDza/
KinDzaDza/kin_dza_dza_1986_part1.avi
KinDzaDza/kin_dza_dza_1986_part1.srt
KinDzaDza/kin_dza_dza_1986_part2.avi
KinDzaDza/kin_dza_dza_1986_part2.srt
total: matches=0  hash_hits=0  false_alarms=0 data=1473476623

sent 1473476623 bytes  received 92 bytes  19261133.53 bytes/sec
total size is 1473296482  speedup is 1.00

Comment 3 Christoph Wickert 2011-07-30 20:22:07 UTC
This is not a problem with grsync because grsync just calls an rsync command line.

This being said when you test grsync vs. rsync, make sure to use exactly the same options. In grsync use 'File -> rsync command line' to figure out the exact command line and then run it on a terminal. Needless to say use exactly the same files.

As for the kernel crash it should be in /var/spool/abrt somewhere. Use 'abrt-cli --list' or abrt-gui to find out.

I am going to reassign this to rsync itself because it is not caused by grsync. And it doesn't have anything to do with LXDE either, so I adjusted the summary of the bug.

Comment 4 nomnex 2011-07-31 04:46:09 UTC
Created attachment 515994 [details]
phone photo of the message after kernel crash

According the shell message after the crash, there was an abrt report about the issue, but it was nothing in /var/spool/abrt (only some reports about pyhook at the give date). I have looked in /var/log/messages (nothing there either).

the only info I have at hand is a print screen I took from my phone when the crash occurred (see attachment).

I will try exactly the same command with a same set of files, along next week, as you requested, and post back.

Comment 5 Vojtech Vitek 2011-08-08 11:22:07 UTC
Also try to run rsync in verbose mode (rsync -vvv), please.

Seems like you were using EXT4 fs on USB hdd, am I right? What fs was used on the PC hdd?

Comment 6 nomnex 2011-08-09 08:39:21 UTC
Hi, I did not have time to try a new backup with grsync vs. rsync yet (I am not very handy with the command line and rsync options). I will use the -vvv mode.

You are correct, ext4 on the HDD usb and ext4 on the PC, both disks share the same user, permission, etc.

Comment 7 nomnex 2011-08-17 13:44:17 UTC
I have tried some massive backups PC > USB HDD using rsync/grsync with the same perm. sets and the same files (~100 GB back and forth), and so far, I can't reproduce the bug. There's been no crash. Grsync log output the same as rsync verbose command, speed is similar. I can't say why I've experienced a Kernel panic doing a rsync backup, but I hope it was a one time happening.

Comment 8 Vojtech Vitek 2011-08-22 10:25:33 UTC
We don't have enough information, so I can't even say if it was bug in rsync or ext4 kernel code.

Imho, the below messages more points to the kernel code..

> rsync: writefd_unbuffered failed to write 4 bytes to socket [sender]: Broken
> pipe (32)
> rsync: write failed on "/media/disk2/torrent/avi/0004.avi": Read-only file
> system (30)
> rsync error: error in file IO (code 11) at receiver.c(322) [receiver=3.0.8]
> rsync: connection unexpectedly closed (70 bytes received so far) [sender]
> rsync error: error in rsync protocol data stream (code 12) at io.c(601)

> kernel BUG at fs/ext4/inode.c:2188!
> invalid opcode: 0000 [#1] SMP
> last sysfs file: /sys/devices/pci000...

Closing INSUFFICIENT_DATA.
Feel free to reopen the bug once you can reproduce the bug again and provide us more information..

Thanks!

Comment 9 nomnex 2011-08-22 13:07:25 UTC
I will if it happens again. Thanks for the feedback.


Note You need to log in before you can comment on or make changes to this bug.