reiserfs-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* How can checksums (node41) be of real help?
@ 2016-05-24  8:50 Georgios Tsalikis
  0 siblings, 0 replies; 3+ messages in thread
From: Georgios Tsalikis @ 2016-05-24  8:50 UTC (permalink / raw)
  To: reiserfs-devel

I suppose they are great for discovering corrupted files, but is there 
any way we can use them to retrieve a good copy of corrupted files?


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: How can checksums (node41) be of real help?
@ 2016-05-24 10:01 Edward Shishkin
  2016-05-24 13:45 ` Edward Shishkin
  0 siblings, 1 reply; 3+ messages in thread
From: Edward Shishkin @ 2016-05-24 10:01 UTC (permalink / raw)
  To: Reiserfs development mailing list

On 05/24/2016 10:50 AM, Georgios Tsalikis wrote:

 > I suppose they are great for discovering corrupted files, but is 
there any way we can use them to retrieve a good copy of corrupted files?

Hello,

Currently if checksum verification failed, then the file system becomes 
readonly
with the suggestion (in kernel messages) to repair the partition with fsck.

Ofc, it is not convenient, but better than nothing. More efficient way 
is to support replicas, so
that every time when checksum verification failed, you upload respective 
replica. It is not difficult
to implement. Just allocate/release replicas simultaneously with 
conventional formatted blocks.
Pointers to replicas can be placed near the checksums (a new node42 
format is needed).
The patches 3.18.6-reiser4-add-node41-support.patch and 
reiser4progs-1.0.9-add-node41-support.patch
are a good example of how to add a new node format.

Thanks,
Edward.



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: How can checksums (node41) be of real help?
  2016-05-24 10:01 Edward Shishkin
@ 2016-05-24 13:45 ` Edward Shishkin
  0 siblings, 0 replies; 3+ messages in thread
From: Edward Shishkin @ 2016-05-24 13:45 UTC (permalink / raw)
  To: Georgios Tsalikis; +Cc: Reiserfs development mailing list

On 05/24/2016 12:01 PM, Edward Shishkin wrote:
> On 05/24/2016 10:50 AM, Georgios Tsalikis wrote:
>
> > I suppose they are great for discovering corrupted files, but is 
> there any way we can use them to retrieve a good copy of corrupted files?
>
> Hello,
>
> Currently if checksum verification failed, then the file system 
> becomes readonly
> with the suggestion (in kernel messages) to repair the partition with 
> fsck.
>
> Ofc, it is not convenient, but better than nothing. More efficient way 
> is to support replicas, so
> that every time when checksum verification failed, you upload 
> respective replica. It is not difficult
> to implement. Just allocate/release replicas simultaneously with 
> conventional formatted blocks.
> Pointers to replicas can be placed near the checksums (a new node42 
> format is needed).
> The patches 3.18.6-reiser4-add-node41-support.patch and 
> reiser4progs-1.0.9-add-node41-support.patch
> are a good example of how to add a new node format.


The cleanest option, however, would be to implement external mirrors in the
context of reiser4 logical volumes support, that I am currently working on.
So if you are patient, then wait for the announce of basic things (I 
think this
summer, but not sure).

Thanks,
Edward.

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2016-05-24 13:45 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-05-24  8:50 How can checksums (node41) be of real help? Georgios Tsalikis
  -- strict thread matches above, loose matches on Subject: below --
2016-05-24 10:01 Edward Shishkin
2016-05-24 13:45 ` Edward Shishkin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).