RN104 Keeps crashing/hanging

Can you try the latest 6.1.9 RC?: http://www.readynas.com/forum/viewtopic.php?f=154&t=72282

In the 6.1.9 RC you should see a new option that allows you to disable sync writes for iSCSI. Can you see if that option helps you?

I'll give it a go, but isn't turning off sync writes a little dangerous when it comes to VMware storage?

The NAS does have a UPS attached but what happens if the NAS crashes while VMWare is writing to the storage?

On another note - How do you get the RN104 to communicate with an APC NMC2 attached to a smart ups? Every time I try to add it as a UPS it just comes back with an comms error.

ANDY,
Some questions.
(a) Is Continuous protection ticked for LUN?
(b) To confirm that you are using 1 NIC on NAS, and not 2. No Teaming?

(c) Re UPS, can you connect to NAS via USB?
(d) If no progress, can you put up a Dropbox link for logs.

Thanks, Marto

Turning off sync writes can help with performance and *might* help you, but as you thought turning it off does put your data at greater risk.

However for using the NAS as a datastore for ESXi you really should be looking at one of our business class offerings.

The 104 is a model targeted at the home users and is not designed to be used for the purpose you are using it for.

Maybe you might get away with using it for 1-2 VMs which aren't used that much, but not for what you're running. See e.g. http://www.readynas.com/forum/viewtopic.php?p=427615#p427615

The models that should have been considered for this would include

Budget option (not sure if this would be up to your load, but it would be better than the 104) - ReadyNAS 300 series
Better option - ReadyNAS 516 or 716
Best - ReadyDATA 516

Also, do you backup the VMs? It is important to backup the VMs regularly.

The VM's in question are not heavily used so don't put much load on the nas storage in terms of IO's. This is set up for a charity on a tight budget hence the reason for choosing the RN104 which is VMWare certified after all. Performance is fine when all is running as it should and the guests respond very quickly. So I don't think this is a performance issue.

We are using both NICs on the RN, but not in a teamed mode. The ESXi cluster is setup to use both storage nics on the hosts and the RN as multipathed LUNS which is the VMWare best practice way of setting it up.

Backups are performed via Veeam to a separate server and then transferred offsite for security. Only one guest is backed up at once so as to not cause any performance issues.

Continuous protection is disabled as is everything else except iscsi

As for the logs, I cleared them down a couple of weeks ago following one of the crashes as there was nothing obvious in there. The RN crashed again last night and this is the sum total of the logs. Just to clarify, when the RN crashes, the web interface is unavailable so the only way to recover is a power cycle.

Wed Jul 30 2014 20:17:47
System: ReadyNASOS background service started.
Wed Jul 30 2014 20:17:41
System: ReadyNASOS service or process was restarted


The UPS cannot be connected via USB as it only has a serial port and is also used to shutdown the VMWare hosts via the NMC2. SNMP is enable on the card, but the RN just doesn't want to know.

A 2 node VMWare cluster hosting 5 VMs may sound small to you, but it's a lot for the 102 and probably too much.

You could see if disabling sync writes helps but the problem may reappear.

The status logs don't tell us much. It is the logs zip file that has the info that can help tell why a system crashed. See http://www.readynas.com/kb/faq/misc/how_do_i_send_all_logs

I wasn't aware that downloading the logs gave you the full system logs!

These have been downloaded and emailed as per the instructions. It may be down to performance in which case I may have to migrate some guests off the cluster or buy another NAS to spread the load. I have already upgraded to 6.1.9 rc8 and disabled sync write to see if this helps so will monitor for it happening again.

The UPS issue still exists. When trying to add it via the web interface as a snmp ups, this is the error that I am getting

Failed to communicate with UPS.<br />Error in during post proc

Code: 17003030001

Datetime: 08:16:31 08/02/2014