Can FreeNAS notify me of disk failure? How? SNMP?

[HeepH]

I'm looking for ways to have FreeNAS notify me of (impending) disk doom. I would like to be able to monitor the state of my disks via SNMP using my monitoring tool of choice (WhatsUp Gold).

If there are no SNMP mib's available for FreeNAS, I could always just let it mail me when something goes wrong, but I keep hearing that FreeNAS is relatively "bad" (slow) at detecting disk failure. Therefor I'm afraid it won't detect the failure in a timely manner.

What would be the best way to keep myself up to date with my disks status?

 

[Robert Trevellyan]

I could always just let it mail me when something goes wrong

This is the usual approach. By default SMART data is checked every 30 minutes. You should also schedule regular SMART tests. Then assuming you have email notifications working correctly, you should be all set.

I keep hearing that FreeNAS is relatively "bad" (slow) at detecting disk failure

Really? How would FreeNAS be slower than any other system?

 

[HeepH]

This is the usual approach. By default SMART data is checked every 30 minutes. You should also schedule regular SMART tests. Then assuming you have email notifications working correctly, you should be all set.

Gotcha, though I would prefer integration with WhatsUp Gold, so I can have a nice overview of all my hardware. Any chance for a SNMP solution?

Really? How would FreeNAS be slower than any other system?

Slow may have been a bad word. Lets call it "unaware/ignorant".

During my searches on SNMP (with limited result :( ) statements kept popping up with people observing FreeNAS ignoring or not properly recognizing disk failures / relocated sectors / hot-swapped or removed disks and this kinda "scares" me. Here for instance, albeit an old topic.

 

[Robert Trevellyan]

Any chance for a SNMP solution?

No idea, sorry.

That's a pretty old thread you quoted (2012).

I suggest experimenting with a toy system to see if it behaves according to your expectations.

 

[HeepH]

No idea, sorry.

That's a pretty old thread you quoted (2012).

I suggest experimenting with a toy system to see if it behaves according to your expectations.

Yeah, im in the process of building something in the testlab, but I cant fake bad sectors or re-allocations right? And I'm looking at detecting early signs mostly. Just so that I'm guaranteed not to get caught off-guard. (Raidz2+ hotspare - not too worried, but hey)

 

[danb35]

I don't have any use for SNMP, and consequently don't know much, if anything about it, but FreeNAS does have some support for it--see the manual here: http://doc.freenas.org/9.3/freenas_services.html#snmp

As for email notifications, there are two basic types that can happen. First, if a disk simply drops offline, or otherwise fails catastrophically, FreeNAS will know about it right away. If you have email set up properly, you'll get an email notification within a minute or so noting the problem.

For less-critical issues, if you've enabled the SMART service, FreeNAS scans the SMART data on your disks at the interval you specify. If it finds a problem, like one or more bad sectors, or a drive that's too hot, it emails you at the address you configured for SMART alerts. This doesn't happen quite as quickly (I think the default SMART scan interval is 30 minutes), but these issues are not as critical.

 

[Robert Trevellyan]

I'm looking at detecting early signs mostly.

I think you'll be happy with how FreeNAS handles this, assuming you have it set up correctly. The thread you referenced was predicated on testing by pulling a disk, which would simulate catastrophic failure and is less likely to happen without warning.

Raidz2+ hotspare

Why do RAIDZ2 with a hot spare when you can do RAIDZ3? Or will you have multiple vdevs sharing the spare?

 

[HeepH]

Why do RAIDZ2 with a hot spare when you can do RAIDZ3? Or will you have multiple vdevs sharing the spare?

Well, hotspare is not technically the correct term. We just keep some disks in reserve since we have several devices using the same disks. So they are technically ehm, cold spares, I guess? haha!

In any case, thank all of you for your information and assurance. I'm going to run some small tests with email notification and judging by your comments that should suffice.