I recently used dar to make a backup of my system, and was given the following error halfway through the process:
Code:
Error while saving /home/myuser/.thunderbird/default.bnf/Mail/News & Blogs/News: Error while reading from file: Input/output error
I emailed the error to dar's support mailing list, and was told to try cat'ing the file to /dev/null and making a backup of just that single file, to see if the problem recurred. I did both (99% sure I typed everything right), and wasn't given any errors.
I then tried the full backup again, and was given the same error again. I tried the cat and single backup again, and this time get IO errors on both:
Code:
> cat /home/myuser/.thunderbird/default.bnf/Mail/News\ \&\ Blogs/News > /dev/null
cat: /home/myuser/.thunderbird/default.bnf/Mail/News & Blogs/News: Input/output error
> dar [just that one file] > /dev/null
Adding file to archive: /home/myuser/.thunderbird/default.bnf/Mail/News & Blogs/News
Error while saving /home/myuser/.thunderbird/default.bnf/Mail/News & Blogs/News: Error while reading from file: Input/output error
Writing archive contents...
--------------------------------------------
0 inode(s) saved
with 0 hard link(s) recorded
0 inode(s) changed at the moment of the backup
0 inode(s) not saved (no file change)
1 inode(s) failed to save (filesystem error)
38 files(s) ignored (excluded by filters)
0 files(s) recorded as deleted from reference backup
--------------------------------------------
Total number of file considered: 39
--------------------------------------------
EA saved for 0 file(s)
--------------------------------------------
I don't like the sound of "filesystem error" very much...
What would cause an IO error like this to happen? Should I worry about general filesystem corruption, or is this just some quirk with that one file? If it matters, running "file" on it gives me:
Code:
.thunderbird/default.bnf/Mail/News & Blogs/News: UTF-8 Unicode English text, with very long lines, with CRLF, LF line terminators
I can also edit the file with vim and view all of it with no problems.
Thanks in advance.
--EDIT--
If it helps, here's the end of the result of running strace on the cat command:
Code:
read(3, "\n padding: 0;\n }\n "..., 4096) = 4096
write(1, "\n padding: 0;\n }\n "..., 4096) = 4096
read(3, 0x804e3b0, 4096) = -1 EIO (Input/output error)
write(2, "cat: ", 5cat: ) = 5
write(2, ".thunderbird/default.bnf/Mail/Ne"..., 47.thunderbird/default.bnf/Mail/N ews & Blogs/News) = 47
open("/usr/share/locale/en_US/LC_MESSAGES/libc.mo", O_RDONLY) = -1 ENOENT (No suc h file or directory)
open("/usr/share/locale/en/LC_MESSAGES/libc.mo", O_RDONLY) = -1 ENOENT (No such f ile or directory)
write(2, ": Input/output error", 20: Input/output error) = 20
write(2, "\n", 1
) = 1
close(3) = 0
close(1) = 0
exit_group(1) = ?
--EDIT 2--
Some more relevant info: I checked /var/log/messages, and I don't really know what I'm looking for, but I don't see anything interesting there. Also, when I try to cat the file without sending it to /dev/null, I get most of the file printed and then the IO error:
Code:
lots of things fly by
...
iframe {
position: fixed;
top: 0;
right: 0;
bottom: 0;
left: 0;
border: none;
}
</style>
</head>
<body>
<iframe id ="_mailrssiframe" src="http://www.cat: News: Input/output error
--EDIT 3--
Sorry for writing this post as I go...
I tried viewing the same part of the file in vim, and I notice the "<body>" tag in the error above is the very last thing that appears in the file in vim, on line 17511.
--EDIT 4--
Well, it looks like about the worst thing that can happen has... I just checked my /var/log/syslogs, and I get the following fun error:
Code:
> grep dma /var/log/syslog*
syslog:Mar 28 00:20:58 mymachine kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
syslog:Mar 28 00:20:58 mymachine kernel: hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=49213032, sector=29088304
syslog:Mar 29 13:05:42 mymachine kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
syslog:Mar 29 13:05:42 mymachine kernel: hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=49213032, sector=29088304
syslog:Mar 29 13:05:46 mymachine kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
syslog:Mar 29 13:05:46 mymachine kernel: hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=49213032, sector=29088312
syslog:Mar 29 14:41:04 mymachine kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
syslog:Mar 29 14:41:04 mymachine kernel: hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=49213032, sector=29088304
syslog:Mar 29 14:41:08 mymachine kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
syslog:Mar 29 14:41:08 mymachine kernel: hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=49213032, sector=29088312
Most things I read online say this means the drive is physically bad and needs to be replaced, although a couple point to filesystem corruption that could be fixed. I've backed up what I can and I'll continue to do that so I make sure nothing gets lost, and I'm wondering what my next step should be. What exactly is this error, and should I try fsck or something else. I have a ReiserFS filesystem.
--EDIT 5--
I just read over the above errors again, and it appears they only appeared when I was making a backup that tried to read that file or when I was doing the testing on it with cat (as opposed to being random or periodic) if that makes any difference.
--EDIT 6--
I just booted up a Live CD to do a little drive checking. I ran reiserfsck --check and was told the filesystem was clean with no problems. I then ran badblocks on the partition which returned with no output, I assume meaning that none were found. What am I to make of this? Do I believe the kernel logs which are indicating drive problems, or believe badblocks, which is supposed to check if they exist? Is there a chance the default badblocks configuration will not detect some bad blocks?
