HAST + ZFS + NFS + CARP

Fri Jul 1 11:09:58 UTC 2016

Am 01.07.2016 um 12:57 schrieb Julien Cigar:
> On Fri, Jul 01, 2016 at 12:18:39PM +0200, InterNetX - Juergen Gotteswinter wrote:
>> Am 01.07.2016 um 12:15 schrieb Julien Cigar:
>>> On Fri, Jul 01, 2016 at 11:42:13AM +0200, InterNetX - Juergen Gotteswinter wrote:
>>>>>
>>>>> Thank you very much for those "advices", it is much appreciated! 
>>>>>
>>>>> I'll definitively go with iSCSI (for which I haven't that much 
>>>>> experience) over HAST.
>>>>
>>>> good luck, i rather cut one of my fingers than using something like this
>>>> in production. but its probably a quick way if one targets to find a new
>>>> opportunity ;)
>>>
>>> why...? I guess iSCSI is slower but should be safer than HAST, no?
>>
>> do your testing, please. even with simulated short network cuts. 10-20
>> secs are way enaugh to give you a picture of what is going to happen
> 
> of course I'll test everything properly :) I don't have the hardware yet
> so ATM I'm just looking for all the possible "candidates", and I'm 
> aware that a redundant storage is not that easy to implement ...
> 
> but what solutions do we have? It's either CARP + ZFS + (HAST|iSCSI), 
> either zfs send|ssh zfs receive as you suggest (but it's
> not realtime), either a distributed FS (which I avoid like the plague..)

zfs send/receive can be nearly realtime.

external jbods with cross cabled sas + commercial cluster solution like
rsf-1. anything else is a fragile construction which begs for desaster.

> 
>>
>>>>
>>>>>
>>>>> Maybe a stupid question but, assuming on the MASTER with ada{0,1} the 
>>>>> local disks and da{0,1} the exported iSCSI disks from the SLAVE, would 
>>>>> you go with:
>>>>>
>>>>> $> zpool create storage mirror /dev/ada0s1 /dev/ada1s1 mirror /dev/da0
>>>>> /dev/da1
>>>>>
>>>>> or rather:
>>>>>
>>>>> $> zpool create storage mirror /dev/ada0s1 /dev/da0 mirror /dev/ada1s1
>>>>> /dev/da1
>>>>>
>>>>> I guess the former is better, but it's just to be sure .. (or maybe it's
>>>>> better to iSCSI export a ZVOL from the SLAVE?)
>>>>>
>>>>
>>>> are you really sure you understand what you trying to do? even if its
>>>> currently so, i bet in a desaster case you will be lost.
>>>>
>>>>
>>>
>>> well this is pretty new to me, but I don't see what could be wrong with:
>>>
>>> $> zpool create storage mirror /dev/ada0s1 /dev/da0 mirror /dev/ada1s1
>>> /dev/da1
>>>
>>> Let's take some use-cases:
>>> - MASTER and SLAVE are alive, the data is "replicated" on both
>>>   nodes. As iSCSI is used, ZFS will see all the details of the
>>>   underlying disks and we can be sure that no corruption will occur
>>>   (contrary to HAST)
>>> - SLAVE die, correct me if I'm wrong the but pool is still available,
>>>   fix the SLAVE, resilver and that's it ..?
>>> - MASTER die, CARP will notice it and SLAVE will take the VIP, the
>>>   failover script will be executed with a $> zpool import -f
>>>
>>>>> Correct me if I'm wrong but, from a safety point of view this setup is 
>>>>> also the safest as you'll get the "fullsync" equivalent mode of HAST
>>>>> (but but it's also the slowest), so I can be 99,99% confident that the
>>>>> pool on the SLAVE will never be corrupted, even in the case where the
>>>>> MASTER suddently die (power outage, etc), and that a zpool import -f
>>>>> storage will always work?
>>>>
>>>> 99,99% ? optimistic, very optimistic.
>>>
>>> the only situation where corruption could occur is some sort of network
>>> corruption (bug in the driver, broken network card, etc), or a bug in
>>> ZFS ... but you'll have the same with a zfs send|ssh zfs receive
>>>
>>>>
>>
>> optimistic
>>
>>>> we are playing with recovery of a test pool which has been imported on
>>>> two nodes at the same time. looks pretty messy
>>>>
>>>>>
>>>>> One last thing: this "storage" pool will be exported through NFS on the 
>>>>> clients, and when a failover occur they should, in theory, not notice
>>>>> it. I know that it's pretty hypothetical but I wondered if pfsync could
>>>>> play a role in this area (active connections)..?
>>>>>
>>>>
>>>> they will notice, and they will stuck or worse (reboot)
>>>
>>> this is something that should be properly tested I agree..
>>>
>>
>> do your testing, and keep your clients under load while testing. do
>> writes onto the nfs mounts and then cut. you will be surprised about the
>> impact.
>>
>>>>
>>
>>
>>
>>>>> Thanks!
>>>>> Julien
>>>>>
>>>>>>
>>>>>>>>>> ZFS would then know as soon as a disk is failing.
>>>>>>>>>> And if the master fails, you only have to import (-f certainly, in case of a master power failure) on the slave.
>>>>>>>>>>
>>>>>>>>>> Ben
>>>>>> _______________________________________________
>>>>>> freebsd-fs at freebsd.org mailing list
>>>>>> https://lists.freebsd.org/mailman/listinfo/freebsd-fs
>>>>>> To unsubscribe, send any mail to "freebsd-fs-unsubscribe at freebsd.org"
>>>>>
>>>
>>
> 