25.09.2014 21:58, Leyne, Sean wrote:
> Your posts suggest that you have fundamental (almost philosophical) problems
> with the idea of using SMB v3 storage for FB purposes.
I have one fundamental problem indeed: I know what clusters can do and what
they
cannot. Unfortunately, they have no m
On 9/25/2014 3:43 PM, Dimitry Sibiryakov wrote:
> 25.09.2014 21:27, Leyne, Sean wrote:
>> Installing on physical server means that you have a single point of failure.
> Modern storage systems claim that they have no single point of failure,
> BTW. Some of
> them even allows two or more servers
>Modern storage systems claim that they have no single point of failure,
> BTW. Some of them even allows two or more servers to be attached.
I know, I have bought a couple of those systems in the past.
>But all this is pointless because Firebird server itself is a SPoF.
True, but that
> 25.09.2014 21:58, Leyne, Sean wrote:
> > All changes are written to disk first
>
>Which one in, say, ten nodes cluster?
A cluster share always has an elected/nominated primary node (set per
share/resource). The non-primary nodes can service all requests but write
requests are re-direct
25.09.2014 21:58, Leyne, Sean wrote:
> All changes are written to disk first
Which one in, say, ten nodes cluster?
> Further, all that is required for a write operation is to send a message
> which invalidates the cache object when the write happens. The propogation
> of the changed data ca
> 25.09.2014 21:27, Leyne, Sean wrote:
> > Having a SQL/FB running in VM on a HA VM cluster connected to HA disk
> storage cluster provides a high level of resiliency, with multiple failures
> necessary in order for functionality to be lost.
>
>I wonder how they manage synchronization betwee
25.09.2014 21:27, Leyne, Sean wrote:
> Installing on physical server means that you have a single point of failure.
Modern storage systems claim that they have no single point of failure, BTW.
Some of
them even allows two or more servers to be attached.
But all this is pointless because Fi
25.09.2014 21:27, Leyne, Sean wrote:
> Having a SQL/FB running in VM on a HA VM cluster connected to HA disk storage
> cluster provides a high level of resiliency, with multiple failures necessary
> in order for functionality to be lost.
I wonder how they manage synchronization between nodes
> 25.09.2014 18:25, Leyne, Sean wrote:
> > I am not aware of SQL Server running on SMB anything other than SMB v3.
> It would be far too dangerous.
>
>If a server supports SMB3, it means that it has Windows inside, because
> Samba doesn't support SMB3 yet. So, I wonder what can preventing an
25.09.2014 18:25, Leyne, Sean wrote:
> I am not aware of SQL Server running on SMB anything other than SMB v3. It
> would be far too dangerous.
If a server supports SMB3, it means that it has Windows inside, because
Samba doesn't
support SMB3 yet. So, I wonder what can preventing anyone fro
> > Jiri Could you do some tests on a smb 3 share with Firebird ?
>
> I don't have a device with SMB3 support or Windows 2012 machine. I can
> install it to VM, but that's not going to show valuable data, I think.
To truly test SMB 3 you would need to create a MS cluster, which will require a
> > I do not see SMB 2 as an appropriate for use for database functions.
>
> I was confused by SMB2/2.1 as well. I thought it might be good to investigate
> what they are doing (preventing problems, how the share is opened, ...).
I am not aware of SQL Server running on SMB anything other than S
> Jiri Could you do some tests on a smb 3 share with Firebird ?
I don't have a device with SMB3 support or Windows 2012 machine. I can install
it to VM, but that's not going to show valuable data, I think.
--
Mgr. Jiří Činčura
Independent IT Specialist
> I do not see SMB 2 as an appropriate for use for database functions.
I was confused by SMB2/2.1 as well. I thought it might be good to investigate
what they are doing (preventing problems, how the share is opened, ...).
--
Mgr. Jiří Činčura
Independent IT Specialist
`
> The why is explained by users , There are a lot of nas/raid devices that can
> be
> used for db storage also for clustering (cold standby)
>
> Jiri Could you do some tests on a smb 3 share with Firebird ?
>
> http://tracker.firebirdsql.org/browse/CORE-3915
>
> On Wed, Sep 24, 2014 at 12:00
24.09.2014 14:58, marius adrian popa wrote:
> The why is explained by users , There are a lot of nas/raid devices that can
> be used for
> db storage also for clustering (cold standby)
I bet that it is much easier to find NAS which support iSCSI or NFS. Do you
know a
popular SMB-only NAS dev
The why is explained by users , There are a lot of nas/raid devices that
can be used for db storage also for clustering (cold standby)
Jiri Could you do some tests on a smb 3 share with Firebird ?
http://tracker.firebirdsql.org/browse/CORE-3915
On Wed, Sep 24, 2014 at 12:00 PM, Dimitry Sibiryako
24.09.2014 10:44, Jiří Činčura wrote:
> Maybe we can investigate how are they doing it and how reliable it is and
> what are the gotchas.
For the beginning it would be enough "what it is good for". The post is
describing
"how" but say no word about "why".
--
WBR, SD.
--
http://blogs.technet.com/b/josebda/archive/2011/02/24/sql-over-smb2-one-of-the-top-10-hidden-gems-in-sql-server-2008-r2.aspx
Maybe we can investigate how are they doing it and how reliable it is and what
are the gotchas.
--
Mgr. Jiří Činčura
Independent IT Specialist
--
19 matches
Mail list logo
