mjyounger

Creation of High Availability Cluster failure

Recommended Posts

Posted (edited)

Hi All,

 

First time posting. But long term user of DSM on Synology hardware - don't shoot me! ;-)

 

So, I have a synology HA cluster of (obvs) two Synology NAS' and following a recent false diagnosis of yet another hardware failure (by Synology support) of the failure of the mainboard on one of my NAS's I am wanting to build my own higher spec ones from commodity PC parts so that I am in control of the hardware quality and maintenance if a part fails without having to replace the while blasted thing. To which end I have just spun up two 3617xs's with identical virtual hardware on ESXi 6.7 having followed all the instructions etc with grub.cfg and the boot loader.

 

DSM: 6.2.2-24922 u6

Synoboot.img : ds3617_6.2

DS3615.vmx : DS3615xs 6.0.2 Jun's Mod V1.01

 

Being new to this I am unsure if there is any other missing info....apologies.

 

All functioning fine until I try and setup HA. It gets right to the end of the process (passes all the tests / checks etc) at which point (as far as I can tell) It sends the config to the passive node and asks it to reboot - at which point it waits and waits and waits for node 2 to reboot but it does not! 

 

I have tried a VM reset but that cuts the power and powers it on a-fresh which seems to break the chain of whatever is going on - presumably there is some config held somewhere which is lost....the screen does say not to power off! Lol - I was desperate.

 

I didn't know if there was a trick to getting the DSM to be able to reboot itself (which seems to work fine if I do it from the GUI myself or indeed from the SSH console) in order that I can then complete the HA Cluster setup.

 

I should add that the reason I am doing this is just some testing prior to buying the hardware and setting it all up on physical kit. so if anyones experience would suggest that this is a nuance of virtual DSM then I am pretty happy to explain it away as that and keep my fingers crossed for the real deal in the physical world.

 

Finally, slightly off topic, but if I have an (well two!)  8 disk array(s) on my "real" Synology DSM nodes and literally pull the disks and plug them in to my new home-brew xpenology, as long as I can ensure that the disks keep the same order (presumably) on the Raid controller.... should I be reasonably confident that my data will just remain intact in my data volume and perhaps just the system volume (which DSM is on) will need a reinstall / overwrite with xpenology? or is the reality not that simple....

 

thanks all!

 

Mark.

Edited by mjyounger
added version info

Share this post


Link to post
Share on other sites

I've only used HA mode on real Syno kit so no direct experience of this with XPE, but I'd look at the following;

 

Serial numbers and MAC addresses being 'genuine' (bearing in mind that the XPE community discourages using genuine numbers on XPE)

S/N and MACs aside,

Try a test with and XPE 'Master' and real Syno 'slave' to see if it works

Then vice versa.

Might help eliminate if there is a hardware issue

Check the system logs in either case

Depending on your hypervisor lab setup, are your VM's HA NICs on the same physical NIC or separate? There might be an underlying issue with that so try different NICs

I think the best way to prove this would be physical kit rather than VMs, even older hardware if you have spare kit, before you buy new.

Also try the 3615 boot loader and lower boot versions.

 

About moving between real syno and XPE, provided you have fully tested your XPE hardware then it should work, however if your real DSM version is higher than your XPE boot build you might need to 'migrate' and reinstall DSM .pat file and then update to the version you have. If its a major build difference you  might need to force a downgrade by editing various config files or creating a clean install HDD and adding your array drives afterwards and repairing the system partitions. All this presumes you have good, verified data backups of course!

 

Which, for me prompts another question. If you are currently running an HA real Syno environment, I'm guessing this is business and critical environment, so should you be considering using a 'hacked' O/S like XPE that is community supported and could expire at any time, should for example, Synology lock down the O/S. 

Share this post


Link to post
Share on other sites

Hi sbv3000,

 

Firstly, thanks for all the pointers - and I take them all on board and will have a play - having said that I'm not keen on breaking my real SYNO HA Cluster, as the re-sync takes easily a weekend if not more like 3 or 4 days, the other issue is that I have two DS 1817+ real SYNO's and I don't see anything available (boot loaders) on XPENOLOGY which looks close enough that the HA wizard would accept it as "compatible hardware", I guess a 3615xs setup with a storage pool of 8 identical drives might work, but I think its likely it wouldn't... will have to have a think about that.

 

There seem to be some very thorough checks that go in to place prior to the HA wizard making things happen and committing configurations, so pending any nuggets of useful info in the logs - I would be surprised if it is the NICs - for example, the setup wizard passes the HEARTBEAT test... as it happens I have the NIC 1 on each VM attached to my default VM Switch and then a SYN-SYNC switch with an appropriate port group assigned to ONLY Nic 2 on each Xpenology node VM which is effectively a private link - which like I say passes the heartbeat test so the assumption is its sufficient for the purpose.

 

This is a totally "home based" environment, none of what you might call business critical, just happy life critical ;-)... the reason behind all of this for me is that I cannot find any other NAS software which implements (what I would call) real HA in terms of real time replication of data and failover when something happens...the only other one which I have seen thus far was open filer which is still downloadable but O L D and ceased active development (I think) around 2015. everyone else's HA is more like occasional/upto hourly snapshot replication or similar which might work, but just doesn't seem as seamless as the SYNO HA.. My reasons for looking at Xpenology are entirely based upon my personal view of the quality of Synologies hardware and the fact that their mainboards and PSU's are prone to failure (plenty of evidence of this both first hand and anecdotally from the forums) and whilst the mainboard isn't replaceable, lthe ast failure of the PSU I had yielded some significant over expenditure for what is just a relatively Low powered seemingly commodity PSU.  Thus if I can maintain the same, what I consider to be, awesome DSM software but on some easily maintainable PC hardware and at the same time benefit from some more powerful hardware too then this feels like a win win. Also I reckon for the price of a single replacement 1817+ (well - I think now it would have to be an 1819+) I can build two well specced white label NAS'.

 

Finally, I get your point about SYNOLOGY locking down their DSM s/w or doing something to thwart the good people of Xpenology... if that day comes I will just have to suck it up and make FreeNAS or similar work......with that said If anyone knows of any other NAS s/w out there either free or reasonably commercially available and priced, that does true HA Clustering of nodes+storage then I would love to hear from you please!

 

cheers!

 

Mark.

Share this post


Link to post
Share on other sites

Sorry meant to ask.... "check the logs" : just from the DSM gui or are there other possibly more verbose logs availably via SSH'ing in or hidden away somewhere?

Share this post


Link to post
Share on other sites

Your loader version doesn't make sense.  1.01?  Should be 1.03b for DS3615xs and DSM 6.2.  Are you using someone's prebuilt VM?

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.