Hello,
I have a customer who want two nas in HA.
Actually we had to use two TVS-871U-RP because they ask for those models.
But i saw the RTRR is not really in real time and there's no possibility to use a Virtual IP.
GlusterFS is still in wait for qnap announcement ? viewtopic.php?f=121&t=75595
With other hardware we used to configure storage cluster with drbd or gluster + samba + virtualip like that it's transparent for all users.
Do you have a solution to have a REAL HA ?
I tought about a VM in HA on proxmox cluster who manage the replication and the file sharing but this is garbage for network perfs and architecture.
Also tought about a round robin dns and built-in RTRR but i think it can lead to file conflicts when two users open the same file.
If nothing is possible with the qnap system, is it possible to install an other linux distro on this hardware ?
Thank you
Real storage cluster HA
- stefano.pederzani
- Getting the hang of things
- Posts: 99
- Joined: Thu Nov 03, 2016 9:28 pm
- Location: Italy
- Contact:
Re: Real storage cluster HA
Without having a real clustered FS experience, I would try DRBD, for simplicity.
But it is a pain a neck, having a customer that exiges that model of storage and leave in your hands the solution.
But it is a pain a neck, having a customer that exiges that model of storage and leave in your hands the solution.
System Administrator & DBA since 1995
- Trexx
- Ask me anything
- Posts: 5388
- Joined: Sat Oct 01, 2011 7:50 am
- Location: Minnesota
Re: Real storage cluster HA
Something like this is really what you should be looking: ES1640dc
With a dual-controller head, and proper redundancy designed into the raid design, this should give you very high-availability with minimal down-time or data loss. Now whether I would recommend QNAP for that kind of usecase vs say a NetApp or Dell/EMC is a different discussion. But if they TRULY need that kind of availability and data resiliency, it cost $$$.
Trying to "bolt-on" a solution that the vendor doesn't natively support is just asking for headaches.
With a dual-controller head, and proper redundancy designed into the raid design, this should give you very high-availability with minimal down-time or data loss. Now whether I would recommend QNAP for that kind of usecase vs say a NetApp or Dell/EMC is a different discussion. But if they TRULY need that kind of availability and data resiliency, it cost $$$.
Trying to "bolt-on" a solution that the vendor doesn't natively support is just asking for headaches.
Paul
Model: TS-877-1600 FW: 4.5.3.x
QTS (SSD): [RAID-1] 2 x 1TB WD Blue m.2's
Data (HDD): [RAID-5] 6 x 3TB HGST DeskStar
VMs (SSD): [RAID-1] 2 x1TB SK Hynix Gold
Ext. (HDD): TR-004 [Raid-5] 4 x 4TB HGST Ultastor
RAM: Kingston HyperX Fury 64GB DDR4-2666
UPS: CP AVR1350
Model:TVS-673 32GB & TS-228a Offline[/color]
-----------------------------------------------------------------------------------------------------------------------------------------
2018 Plex NAS Compatibility Guide | QNAP Plex FAQ | Moogle's QNAP Faq
Model: TS-877-1600 FW: 4.5.3.x
QTS (SSD): [RAID-1] 2 x 1TB WD Blue m.2's
Data (HDD): [RAID-5] 6 x 3TB HGST DeskStar
VMs (SSD): [RAID-1] 2 x1TB SK Hynix Gold
Ext. (HDD): TR-004 [Raid-5] 4 x 4TB HGST Ultastor
RAM: Kingston HyperX Fury 64GB DDR4-2666
UPS: CP AVR1350
Model:TVS-673 32GB & TS-228a Offline[/color]
-----------------------------------------------------------------------------------------------------------------------------------------
2018 Plex NAS Compatibility Guide | QNAP Plex FAQ | Moogle's QNAP Faq
- stefano.pederzani
- Getting the hang of things
- Posts: 99
- Joined: Thu Nov 03, 2016 9:28 pm
- Location: Italy
- Contact:
Re: Real storage cluster HA
I agree at 100%!Trexx wrote:Something like this is really what you should be looking: ES1640dc
With a dual-controller head, and proper redundancy designed into the raid design, this should give you very high-availability with minimal down-time or data loss. Now whether I would recommend QNAP for that kind of usecase vs say a NetApp or Dell/EMC is a different discussion. But if they TRULY need that kind of availability and data resiliency, it cost $$$.
Trying to "bolt-on" a solution that the vendor doesn't natively support is just asking for headaches.
(I do not know this product, of course, but the concept)
I personally prefer single solutions with redundancy inside. They statistically work better and without "headaches".
Let me say this:
if a customer wants a starting point, the company can give a finish solution;
if a customer needs a solution, the company can choose a product to achieve this.
But if a customer want to establish start and end, and wants the company gives the middle, well... It is not a real thing.
How to say it to them?
One method can be asking thousands and thousands of dollars, until the customer gives up.
System Administrator & DBA since 1995
- storageman
- Ask me anything
- Posts: 5506
- Joined: Thu Sep 22, 2011 10:57 pm
Re: Real storage cluster HA
What does this flim flam wording even mean?stefano.pederzani wrote:Trexx wrote: Let me say this:
if a customer wants a starting point, the company can give a finish solution;
if a customer needs a solution, the company can choose a product to achieve this.
But if a customer want to establish start and end, and wants the company gives the middle, well... It is not a real thing.
How to say it to them?
One method can be asking thousands and thousands of dollars, until the customer gives up.
-
- First post
- Posts: 1
- Joined: Fri Jan 20, 2017 10:54 pm
Re: Real storage cluster HA
That is a good strategy. But what if the customer says he quits?stefano.pederzani wrote:I agree at 100%!Trexx wrote:Something like this is really what you should be looking: ES1640dc
With a dual-controller head, and proper redundancy designed into the raid design, this should give you very high-availability with minimal down-time or data loss. Now whether I would recommend QNAP for that kind of usecase vs say a NetApp or Dell/EMC is a different discussion. But if they TRULY need that kind of availability and data resiliency, it cost $$$.
Trying to "bolt-on" a solution that the vendor doesn't natively support is just asking for headaches.
(I do not know this product, of course, but the concept)
I personally prefer single solutions with redundancy inside. They statistically work better and without "headaches".
Let me say this:
if a customer wants a starting point, the company can give a finish solution;
if a customer needs a solution, the company can choose a product to achieve this.
But if a customer want to establish start and end, and wants the company gives the middle, well... It is not a real thing.
How to say it to them?
One method can be asking thousands and thousands of dollars, until the customer gives up.