Menu

#177 Compare extended attributes with isolated catalog

2.5.0
closed
None
Implemented
5
2018-12-16
2015-12-17
xloem
No

From manpage -C:

no comparison can be performed concerning EA or FSA even if each of them have their own CRC in the catalogue because different ordering as provided by the OS of the items composing EA and FSA may lead the CRC to be different while the EA or FSA are exactly the same, so CRC here is used only to dectect archive corruption.

I think this is an issue -- catalog should compare EA and FSA; otherwise the files differ from the backup. Solutions: include EA and FSA in catalog, or sort them in the same order each time before calculating a CRC.

Discussion

  • Denis Corbin

    Denis Corbin - 2015-12-17

    You took the text out of its context:

    this sentence does not apply to archive comparison which does compare EA and FSA but applies to comparison based on an isolated catalogue, which by nature does not contain data, EA, not FSA (else it would be the same as an normal archive).

    To sort EA in order implies storing them in memory before writing down them into the archive, EA might be huge, (file forks for example under MAC OS X), so this is not feasable.

    If you want to know whether file differs or not from an archive, CRC comparison is not a proof, just a strong probability they are the same. In consequence, comparison based on isolated catalogue is just an approximation. Real comparison is to be done using the real archive in place of the catalogue.

    Cheers,
    Denis

     
  • Denis Corbin

    Denis Corbin - 2015-12-17
    • status: unread --> open
    • assigned_to: Denis Corbin
     
  • xloem

    xloem - 2015-12-18

    I am interested in verifying my files without having to store copies of all of them -- but also being able to back them up when I have the space. If I were interested in writing the code, would dar be the tool for this job, or should I look elsewhere?

    I can see a couple solutions to the huge EA situation: only storing enough of the EA in memory to do the sorting, reading from the disk to determine the sort order, storing a checksum of each EA rather than a checksum of all of them together, or requiring large memory use if the option is enabled.

     
  • xloem

    xloem - 2015-12-18

    For example, you could CRC each EA separately, then sort the CRCs.

     
  • Denis Corbin

    Denis Corbin - 2017-12-31

    I hope you have noted this feature was implemented since release 2.5.0 of October 2015.
    Sorry for the delay updating this feature request!

     
  • Denis Corbin

    Denis Corbin - 2017-12-31
    • Progression: requested --> Implemented
     
  • Denis Corbin

    Denis Corbin - 2017-12-31
    • km stone :): none --> 2.5.0
     
  • Cálestyo

    Cálestyo - 2018-08-07

    @Denis ... this and several other feature request that are already marked as implemented,... are still marked as open as well... perhaps it would be helpful to mark them all as closed :-)

     
  • Denis Corbin

    Denis Corbin - 2018-12-16

    yes correct. Closing this case

     
  • Denis Corbin

    Denis Corbin - 2018-12-16
    • status: open --> closed
     

Log in to post a comment.

MongoDB Logo MongoDB