3

While copying/downloading a large file (3+ GB) the checksusms aren't the same afterwards.

Methods attempted:

  • A patch cable transfer between two computers. (source sum was OK)
  • Download from Main site.
  • Copy from another partition. (the original sum was OK)

The file is an ISO for a Debian-Based OS. (but this is not the only file that has had problems...)

I am using Ubuntu 15.10 alongside Mac OSX El Capitian, and have had this problem on both. However with the Mac I was eventually able to get the file correctly.

I have run the badblocks command (on entire disk) and fsck.ext4 (Only on the Ububntu partition) from a live CD and no problems were reported. More over the SMARTS report for the drive reports the disk is OK.

Could this be a drive failure, and is there anything else I can do to check the integrity of the Hard Drive?

UPDATE:

So another intersting thing is going on... I ran the sha1sum once and got: 87004a6bea1ff1792f733f9048b3c43b7a2c66af I just ran it three more times and got 6607f5a66aee980a8958bb99f9871158a852a319 91b8a6674b9a90ab07c151cedc203a2d9806193f and 076d303bbbce719edffd9ed821bc83c448d42bb4. I have made no changes to this file... Any ideas?

UPDATE 2:

I just restarted my system and the sum has stopped "shifting"... So it appears I might have cut it off mid write, even though cp returned... I am confused at this point as to why it is hanging out in a buffer for so long. Anyone know what I can do about this?

3

I'm guessing bad RAM - it can cause small errors when files are copied, and can be very intermittent, if it's only a small spot of ram that's not always used. I'd expect to see mention of disk read errors in dmesg or /var/log/syslog.

Check with memtest for a few passes, when the computer's "warmed up." I've read that running it for 24hrs is a 100% reliable test, but just a few minutes would find any big errors, an hour should be good.


If the files fit in RAM, they could be only read from disk once, and subsequent reads are from the disk cache in RAM. To clear/flush the disk cache and try another read from disk, run:

sync; echo 3 | sudo tee /proc/sys/vm/drop_caches

Echoing 3 clears all caches, echoing 1 only clears the pagecache (disk caches & other caches I think), 2 frees reclaimable slab objects (includes dentries and inodes) - See Kernel Documentation for /proc/sys/vm/*

  • Thanks for the great answer. It was one of the sticks of RAM. – Gamerb Dec 20 '15 at 2:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.