<html style="direction: ltr;">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<style id="bidiui-paragraph-margins" type="text/css">body p { margin-bottom: 0cm; margin-top: 0pt; } </style>
</head>
<body bidimailui-charset-is-forced="true" style="direction: ltr;">
<pre class="moz-quote-pre" wrap="">* Not a ZFS or encryption expert by any means *
ZFS checks data integrity on every read, and periodically -user defined- can verify the hole file system. When there is a bad read it can correct it from a good copy, if one is available. You can have
a zero redundancy setup, thous you only get the read error and no way of fixing it. You can have more redundancy with more hard disk drives. See ZFS RAID levels.
The send option just lets you have an efficient way to move the data to another box. If there is a read error on the sending box it can be corrected only if another copy is available. ZFS Encryption is at rest: the checks are on the encrypted bits, on the drive, and that is what is sent. You can encrypt the connection -"over the wire"- with an SSH or VPN connection, useful for meta-data and un-encrypted files.
HTH!
On 2/10/22 10:11, Boruch Baum wrote:
</pre>
<blockquote type="cite"
cite="mid:20220210081157.muyogmxsggf4ilqb@E15-2016.optimum.net">
<pre class="moz-quote-pre" wrap="">Can anyone on-list answer two related ZFS questions?
1) block checksum data in 'sent' snapshots
One of the attractive features of ZFS is that it offers a form of
"error correction". My understanding is that what ZFS does actually
is checksum its blocks and upon noticing that one has gone bad,
checks for a copy *somewhere*, where *somewhere* is documented as a
RAID mirror.
I'm considering using ZFS for SSD elements of a home network, for
which I don't see justification for setting up an NFS or RAID mirror
array. Is there an alternative technique to benefit from ZFS error
correction?
The idea that occurs to me is to perform periodic backups using ZFS
send and then to immediately copy the 'sent' snapshots to different
media. My thought here is that if the 'sent' data includes the block
checksums, then should I ever need to perform a restore and be
confronted by a bad block, I could mount the copy as a mirror and
have ZFS perform the correction.
Is such the case? Is this possible? Is there a 'better' technique?
2) native encryption of 'sent' snapshots
For a ZFS pool with native encryption, are its 'sent' data also
natively encrypted? If not, are there any known problems or issues
related to piping 'sent' data through gpg or other encryptor?
--
hkp://keys.gnupg.net
CA45 09B5 5351 7C11 A9D1 7286 0036 9E45 1595 8BC0
_______________________________________________
Linux-il mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Linux-il@cs.huji.ac.il">Linux-il@cs.huji.ac.il</a>
<a class="moz-txt-link-freetext" href="http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il">http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il</a>
</pre>
</blockquote>
</body>
</html>