On Mon, Nov 30, 2020 at 03:12:41 -0500, Gene Heskett wrote:
> Doing a level 0 on /home/gene/Pictures, it logged this in the email:
>
> coyote /home/gene/Pictures lev 0 partial taper: source server crc
> (44cff778:11146117120) and input server crc (dfd0e83a:11146117120)
> differ)
> coyote /home/gene/Pictures lev 0 was successfully retried
>
> It did leave a 10+ Gb file in the vtape, but left the failed files in the
> holding disk:
>
> root@coyote:~$ ls -l /sdb/dumps/20201130020105/
> total 10885096
> -rw------- 1 amanda amanda 2097152000 Nov 30 02:06
> coyote._home_gene_Pictures.0
Would /home/gene/Pictures have changed any between the two retries? If
not, you might learn something by comparing the components of the
successful dump with the files on the holding disk... (but off hand I'm
not sure how many red-herring differences you'd have sort through to
find any hints as to the actual problem).
>
> Does anyone have a clue what its really trying to tell me?
I only have some vague clues:
* the number after the ":" is the size of the file being CRCed. In this
case 11146117120 shows up for both sides of the commparison, so it
seems like the full file got transfered across to whatever step is
causing the error. It also seems like this error applies to the
entire 11GB dump rather than the individual 2GB parts.
* The message "source server crc ([...]) and input server crc" appears
to be generated in Amanda/Taper/Worker.pm:result_cb() in cases where
$self->{'server_crc'} and $self->{'source_server_crc'} differ.
$self->{'server_crc'} seems to be read out of the header of the dump
file itself.
$self->{'source_server_crc'} seems to be computed as part of
transfering the file to the taper process, or something like that.
So I guess the next question is where in the multiple stages of the life
of the dump file the CRC missmatch gets introduced....
I assume that the first few lines of the coyote._home_gene_Pictures.0
file is an Amana header (including an XML chunk); can you post that
here?
Also, what do you find when you grep the Amanda debug/log files for
those two CRC values ( 44cff778 and dfd0e83a )?
One other thought: have the reported CRC errors in the past also been
for the dump of the /home/gene/Pictures DLE, or are multiple different
DLEs affected? Is it always level 0 dumps?
Nathan
----------------------------------------------------------------------------
Nathan Stratton Treadway - [email protected] - Mid-Atlantic region
Ray Ontko & Co. - Software consulting services - http://www.ontko.com/
GPG Key: http://www.ontko.com/~nathanst/gpg_key.txt ID: 1023D/ECFB6239
Key fingerprint = 6AD8 485E 20B9 5C71 231C 0C32 15F3 ADCD ECFB 6239
