*New 11.3 series Release:
2019-10-05: XigmaNAS 11.3.0.4.6928 - released, 11.2 series are soon unsupported!

*New 12.0 series Release:
2019-10-05: XigmaNAS 12.0.0.4.6928 - released!

*New 11.2 series Release:
2019-09-23: XigmaNAS 11.2.0.4.6881 - released!

We really need "Your" help on XigmaNAS https://translations.launchpad.net/xigmanas translations. Please help today!

Producing and hosting XigmaNAS costs money. Please consider donating for our project so that we can continue to offer you the best.
We need your support! eg: PAYPAL

HAST config with only 1 node

Highly Available Storage.
Forum rules
Set-Up GuideFAQsForum Rules
Post Reply
belanzo
NewUser
NewUser
Posts: 2
Joined: 07 Jan 2013 18:13
Location: Hajdúböszörmény - Hungary
Status: Offline

HAST config with only 1 node

#1

Post by belanzo » 21 Jan 2013 11:10

We are planning to create redundant storage with HAST feature but the current budget enough for ony 1 server.
Is it possible to configure CARP+HAST on this machine without the secondary node and later buy and configure it and
after that syncronize the storage for the secondary node?

Best Regards,

BeLANzO

discostur
Starter
Starter
Posts: 17
Joined: 10 Jan 2013 16:24
Status: Offline

Re: HAST config with only 1 node

#2

Post by discostur » 12 Mar 2013 10:50

Hi,

yes this should be possible!

karlandtanya
Starter
Starter
Posts: 45
Joined: 23 Jan 2014 15:31
Location: nelson twp, OH, USA
Status: Offline

Re: HAST config with only 1 node

#3

Post by karlandtanya » 15 Feb 2015 00:25

Me, too--
How would you go about doing this?
I have 2 existing servers with raidz2 filesystems currently backed up by zfs replication.

Can I just pick one node and configure the drives as HAST resources?
At that time telling that (the first) node the configuration *as if* I had already set up the second node?
Can I create the zfs filesystem using hast/ada0...ada5 while the primary node is lonely (other node doesn't exist)?
When I do that will the existing data on the primary node be destroyed?

If it is, can I then copy it over from the other server while there is still no secondary node?
By doing that, HAST will have to keep a *really big* log of "what changed" (which will be everything....)
Note at this point the other server doesn't know anything about HAST and is just a data source.

If I get that far--have one "primary" HAST node that's missing its brother, can I then configure HAST in the other server?
And tell the first one to update the second one?

Is there a risk that the update will go the other way? That is update the server with my data on it by copying the (empty) filesystem from the newly configured "secondary"?

This is my vague plan so far, but I have no intention of pursuing it until a better understanding of HAST!

Yes, I know--the correct answer is buy more servers, set up HAST on them, and THEN copy over all the data.
but this is prohibitively expensive--so I try to shuffle the data around without destroying it!

Any suggestions or advice--even "no--that will not work and here's why" I would very much appreciate!

Thanks!

amason
NewUser
NewUser
Posts: 7
Joined: 19 Dec 2012 20:07
Location: Ottawa, Canada
Status: Offline

Re: HAST config with only 1 node

#4

Post by amason » 16 Jun 2015 19:24

belanzo wrote:We are planning to create redundant storage with HAST feature but the current budget enough for only 1 server.
Is it possible to configure CARP+HAST on this machine without the secondary node and later buy and configure it and
after that synchronize the storage for the secondary node?

Best Regards,

BeLANzO
Yep, just set the Node B IP address to "none"

Then you can create and initialize all HAST resources and they'll be in "degraded" status until you attach the second node.

FYI, when I created a new resource in the GUI yesterday, it didn't work. You can't do a "hastctl role primary X" unless you do a "hastctl create X" first. It fooled me because the resource was listed in "hastctl status", so I'm not sure what's missing.
--
Andy

amason
NewUser
NewUser
Posts: 7
Joined: 19 Dec 2012 20:07
Location: Ottawa, Canada
Status: Offline

Re: HAST config with only 1 node

#5

Post by amason » 16 Jun 2015 19:34

karlandtanya wrote:Me, too--
How would you go about doing this?
I have 2 existing servers with raidz2 filesystems currently backed up by zfs replication.

Can I just pick one node and configure the drives as HAST resources?
At that time telling that (the first) node the configuration *as if* I had already set up the second node?
Can I create the zfs filesystem using hast/ada0...ada5 while the primary node is lonely (other node doesn't exist)?
When I do that will the existing data on the primary node be destroyed?
Yes, see above for the trick. Yes, it'll be very destroyed. You'll need the drive config to be identical (logs, caches, spares included) or there will be manual steps involved on each failover.
karlandtanya wrote: If it is, can I then copy it over from the other server while there is still no secondary node?
By doing that, HAST will have to keep a *really big* log of "what changed" (which will be everything....)
Note at this point the other server doesn't know anything about HAST and is just a data source.

If I get that far--have one "primary" HAST node that's missing its brother, can I then configure HAST in the other server?
And tell the first one to update the second one?
Yes, you should be able to do a ZFS send/receive to transfer all the data across (just noticed you're already doing that, so there's a step saved). I don't think it bothers to keep a log in that case, it just knows that an initial sync will require everything, but either way, it'll work. Yes, step 18 (I skipped a few) will be to configure the HAST secondary to match the primary, and start the sync.
karlandtanya wrote: Is there a risk that the update will go the other way? That is update the server with my data on it by copying the (empty) filesystem from the newly configured "secondary"?

This is my vague plan so far, but I have no intention of pursuing it until a better understanding of HAST!

Yes, I know--the correct answer is buy more servers, set up HAST on them, and THEN copy over all the data.
but this is prohibitively expensive--so I try to shuffle the data around without destroying it!

Any suggestions or advice--even "no--that will not work and here's why" I would very much appreciate!

Thanks!
No, no risk of that, there are clearly defined primary and secondary roles. The main thing is that you want to test a few failure scenarios as part of the setup, so you do want to have lots of scheduled downtime available if you're doing this with live data.

Np, and good luck.
--
Andy

karlandtanya
Starter
Starter
Posts: 45
Joined: 23 Jan 2014 15:31
Location: nelson twp, OH, USA
Status: Offline

Re: HAST config with only 1 node

#6

Post by karlandtanya » 23 Jun 2015 19:47

Sorry I took so long to get back--thanks very much for your advice!

So, this all works, but it turns out that CARP causes problems with Android, so I can't use HAST anyhow.
For now it's regular zfs replication using zrep until I can get past that one!

Post Reply

Return to “HAST”