[OpenIndiana-discuss] [zfs-discuss] zfs incremental send?

[Jonathan Adams]

sure ... but maybe you're lucky ...

whether or not it crashed your system, it definitely _did_ crash ours
... that was on an earlier version of ZFS, but I'm never going to do
it again if I can help it ... we had support calls in with Sun to fix
it ... they couldn't ...

we managed to clone and promote an earlier snapshot of the pool, and
were able import further snapshots into this pool ... but we still had
a zombie snaphot in the chain.

I believe we got it fixed in the end by destroying the "receiving"
pool, sending a USB hard drive to the remote site (to get the large
amount of data sent in the post) and recreating the whole thing again.

After about 5 months Sun got back to us saying that someone else had
had the same problem and that they thought they might have a patch for
it ...

We have procedures in place to do it the long way, I doubt I'll be
prepared to do it the "quick" way again any time soon.

Jon

On 29 March 2011 17:22, Reginald Beardsley <pulaskite at yahoo.com> wrote:
>
> Not sure what you're trying to say.
>
> "zfs send snapshot | zfs receive filesystem"
>
> produced no filesystem when I killed the job prematurely after copying 1+ GB which makes good sense.  Far better not to create a munged filesystem.  Note I tried this w/ S10 U8.
>
> I certainly encountered no problems starting and stopping the job or when I killed it.
>
> I wonder if your problem wasn't the pool "failmode=" setting.  I had a panic w/ zfs 4 when I accidentally dropped power to the USB disk I was writing to.  Rattled me pretty badly.
>
> Have Fun!
> Reg
>
> --- On Tue, 3/29/11, Jonathan Adams <t12nslookup at gmail.com> wrote:
>
>> From: Jonathan Adams <t12nslookup at gmail.com>
>> Subject: Re: [OpenIndiana-discuss] [zfs-discuss] zfs incremental send?
>> To: "Discussion list for OpenIndiana" <openindiana-discuss at openindiana.org>
>> Date: Tuesday, March 29, 2011, 9:52 AM
>> Just to clarify; Our issue was we did
>> a send and receive in the same
>> pipe, when it died the receive died too, and that caused
>> trouble.  the
>> send on it's own was alright, the receive caused problems.
>>
>> _______________________________________________
>> OpenIndiana-discuss mailing list
>> OpenIndiana-discuss at openindiana.org
>> http://openindiana.org/mailman/listinfo/openindiana-discuss
>>
>
>
>
>
> _______________________________________________
> OpenIndiana-discuss mailing list
> OpenIndiana-discuss at openindiana.org
> http://openindiana.org/mailman/listinfo/openindiana-discuss
>