SourceForge has been redesigned. Learn more.
Close

#149 Verification of clone - failure

open-works-for-me
Thomas
5
2014-02-24
2012-05-24
pbea
No

Problem with the following is that I am not sure that it is a Clonezilla 1.2.12-37 issue or mine. My best guess is that it is yours, so here goes. Something on your CD does not like 1tb hard drives, maybe!

I am encountering PartClone image failure messages almost 100% of the time when no protection (fsck) is applied. Each is reproducible, consistently. I power down my box normally and insert the Clonezilla 1.2.12-37 disk and boot up. I begin a partition backup from my 40gb Intel SSD to either of my 1tb WD hard drives at any stage of filling. The image is prepared and saved. The verification fails with (from the /var/log/partclone.log):

"
Partclone v0.2.45 http://partclone.org
Starting to check image (-)
Calculating bitmap... Please wait... done!
File system: EXTFS
Device size: 15.7 GB
Space in use: 6.2 GB
Free Space: 9.5 GB
Block size: 4096 Byte
Used block : 1518420
CRC error again at 0...
"

This error is consistent across all failures. This is 100% reproducible.

I have learned through lots of trial, errors and practice that fsck on the destination 1tb WD EXT4 hard drive 'helps'. I use fsck on my storage devces regularly during the week but this still helps. My routine is to boot Clonezilla to the command prompt and "sudo fsck -f /dev/sdb1" on the destination drive, reboot and backup. I have not started a Clonezilla backup in recent memory without a fsck cleaning. This works 90% of the time.

About 10% of the time the first cycle of fsck doesn't work so I have to cleanse the 1tb WD EXT4 hard drive a second time. I have never had a failure after a second cleaning.

100% of the time I backup a partition to a 110gb ADATA SSD and the image verification is good, each and every time.

I use one 1tb WD EXT4 hard drive at a time to save my work. With daily and hourly incremental backups a drive lasts me a little over a month. With all the activity I can see why fsck may help. I cannot use fsck in a script - too much interactivity. When a drive reaches 90% filled, I go to the other 1tb WD EXT4 hard drive. I use gparted to format the drive to reiserfs. This removes all(?) traces of EXT4. I then reformat the drive as EXT4 and move all my backup scripts over to the new drive. I save a Clonezilla OS partition backup to 3 separate drives at least twice a week.

The 110gb SSD has never had a failure. My earlier reiserfs on the 1tb WD EXT4 hard drives was also hit by this. PartImage and fsarchiver on the SystemRescueCD had no problems. The reason I primarily use your CD is the verificiation. Of my successfully saved Clonezilla backups, not one has failed to restore. Thank You.

This problem has existed with all my versions of Clonezilla. It may be a verification failure bug. I am about to invest in two 3gb hard drives and I need some input as to whether my problems are a result of large drives, ergo the bug report. Short of a low level format (do they do that anymore?) I do not know what else I can do to diagnose the issue.

Sorry for the rambling. Any input is appreciated.

Thank You

Discussion

  • Steven Shiau

    Steven Shiau - 2012-05-25
    • assigned_to: nobody --> thomas_tsai
     
  • Steven Shiau

    Steven Shiau - 2012-05-25

    Please give Clonezilla live 1.2.12-48 a try. It comes with partclone 0.2.48 and the problem might have been fixed.

    Steven.

     
  • pbea

    pbea - 2012-05-25

    Thank You for the quick reply.

    Got a problem. I do not understand your versions. I looked for -48 and I couldn't find it. I did find -52(?) in testing but I am not willing to load testing software for my production machine.

    I did find 'oneiric' which is marked the latest but I am not sure.

    Would you please give me the name and location of what you consider the 'latest stable release' and I will be glad to try it. The version that I am using now, -37 is said on your site to be the latest stable. It will be a few days before I do a full cloning then I will get back to you.

    Hope this helps.

    Thank You.

     
  • pbea

    pbea - 2012-05-30

    I downloaded version -58 and completed a routine backup.

    A 1tb hard drive which is not active but contains about 93% data passed on the first try.

    The current 1tb hard drive which I use on an hourly basis for incremental backups failed the image check with the same log on three "fsck -f /deb/sdb1" tries but then passed on the fourth attempt.

    A 110gb SSD passed on the first try as usual.

     
  • Steven Shiau

    Steven Shiau - 2012-05-30

    Did you also do a memtest check, too?
    Maybe the problem is on the hardware...

    Steven.

     
  • pbea

    pbea - 2012-06-07

    I downloaded -60 and everything is the same - still buggy!

    Do memtest sometimes but did it again - no errors

    Am backed into a corner - workaround must be used - I saved to 120 SSD and copied the good clone to each of the 1tb hard drives - this gives me three backup copies but they are not independently saved

    to recap - I used partimage and fsarchiver on these two 1tb hard drives when they were reiserfs - when I changed to EXT4, I used fsarchiver - no problems - I would continue using fsarchiver but no way of checking copy

    Sent an email to partclone maintainer about using with 1tb hard drives but I never received a reply - I was specifically interested in the 'nochecksum' option - it may be entirely possible that the clone is good and the checksum failed - if I can clone a copy without a checksum check then I can test the result for accuracy - it may just work!

    Hope this helps

     
  • Anonymous

    Anonymous - 2012-06-07

    I wonder about your Cmd line check, you have done it to sdb1.
    Are you sure it is sdb1, you can check that by applying a lable, or using a UUID instead this is directly coupled with your specific devic, while sdx is dependant on your (total) PC hardware?

    A question: you did read /var/log in Terminal. Do you know how to read and save (print) the totaly screen scrolling during Clz execution, especially if in error!

     
  • pbea

    pbea - 2012-06-07

    Thanks for the help

    All my outside and inside storage devices are labelled and I can see lights flashing to confirm the matching label and device.

    I am lost on your last comment what is "during Clz execution" ?

     
  • pbea

    pbea - 2012-06-07

    Sorry to bother you so soon

    I just completed the changeover from one 90% used 1tb hard drive to one newly formatted EXT4 1tb hard drive and I completed a clone job as the first item on the drive.

    100% success - first time - no fsck - no checksum failure

     
  • pbea

    pbea - 2012-06-09

    Good Day . . .

    I tried another cloning and I receive the usual checksum error.

    This bug report is going no place and too much time has been expended so I am reverting back to fsarchiver.

    Keep up the good work and close this bug report.

    Although, keep it in mind should anything like this happen again.

    Thank You

     
  • Steven Shiau

    Steven Shiau - 2012-06-19
    • status: open --> open-works-for-me
     
  • Steven Shiau

    Steven Shiau - 2012-06-19

    Somehow you are the only one reporting this... and it seems we can not reproduce the problem here...
    Feel free to reopen this bug if you have something new.
    Therefore I am closing this bug.

    Steven.

     
  • Joe Beauchamp

    Joe Beauchamp - 2014-02-16

    Ran into the same thing - a killer. Something hit windows server 2008. I don't trust MS for backup and went over to the friend to help - used clonezilla for an independent backup before doing more work in the machine. Every time clonezilla (default settings) would complain about a CRC error on the image. I know there are bad spots on the original, thought I misread the message. Previously ran multiple chkdsks /r on the source partition. No errors on any data, etc. - clean, in other words, for a save. This is 600GB and it takes 1/2 day to try anything. Backup on USB drive. /var/log/partclone.log said the same thing on 3 backups. I took the backup drive, that never had any complaints, in win7 slow format, previously had run chkdsk /r on it, ran chkdsk /r again, of course zero complaints, just a whale of more time gone.
    After the 3rd backup and same complaint, CRC error in image, I took the backup drive over to win7 loaded more files to see if clonezilla deleted the file it called bad (and I would find the truly bad spot by writing something on it and then trying to read it). Read everything on that drive, including what clonezilla put there, with zero problems. 2 days later after all this, I looked around and saw this message, someone else with NO problems on the backup drive yet clonezilla/partclone complained repeatedly about CRC error.
    You know what it is like when every attempt takes 1/2 day. 2 days later, using another product to back up the partition. Another 1/2 day for it, not yet done. Then maybe get back to actually trying to fix the problem? Not nice. I don't need bogus errors. So, do I have a real clonezilla backup or do I just always take my chances with clonezilla? At 1/2 day per try, I can't go on and try other options besides partclone, etc.
    Anyway, it is still a nasty problem. This partition has about 240GB of data/programs. Of course, I can only try 1 thing at a time. Very slow, as many recoveries are. (Clonezilla images come to about 87GB, but maybe that is real since there seem to be tons of log files - MS SQL junk.)

     
  • Steven Shiau

    Steven Shiau - 2014-02-24

    This kind of issue normally is on hardware, either RAM, network cable, or something similar.
    This is very difficult to debug and see why. Therefore I suggest you try different hardware to see if it's always reproducible.

    Steven.

     

Log in to post a comment.