I am considering switching from our current raid 5 solution to unraid. I like a lot of aspects about it and love the community support.

But i have one major concern that is and i can be wrong but it seems unraid has very poor build in error notification.

 

Let me explain:

 

In about 33% of the places i want to use the unraid solution there is limited internet access, a notification mail when a drive fails will be enough but to my surprise unraid cannot do this natively. I know u can add a script, but still.

 

The other 66% has no internet access i can dial in to them and configure the unraid but there is no way to get an email out. So in our current raid solution the raid will beep every 30 secs when a drive failed so the user will notice it and call us, we will then contact them and see what the problem is.

 

Even when i wanted to use this solution privately i would be very nervous about missing a failed drive as i tend not to check the server every day.

 

So how do u monitor effectively the health of the unraid server ?

 

unRAID was designed for a different market, and although it has come a long ways, and there are a number of user addons that may help, I really don't think it is ready for external Internet control.  It was designed (in my opinion) primarily for the local NAS market, home users, which are primarily Windows users, so it fits best on a subnet behind a NAT router.  It is safe then, controlled and monitored locally, through a web page interface.  To fit in a small totally RAM-based image, it is a stripped down Linux distribution, with most of the security stuff removed, ie. no firewall or other security tools and protocols.

 

If you still want to try it, you can add the email notification addon, and teach a local user at each installation how to monitor it, very simple to do.  Other more knowledgeable Linux users here may have better advice, but I would NOT recommend opening these servers up to Internet access for the web based control.

 

Otherwise, you might be better off using a standard Linux distribution, with LVM, or other RAID arrangements.

Ok thank u for your reply, i will look elsewhere for a better solution then. The external control from internet is not really a problem as we connect with vpn to the routers so the unraid server will be protected by the firewall of the router. But sadly it isn't a option to let the user check for emails.

 

A simple beep when in error would be sufficient for our purpose, but i understand now unraid is more orientated to an consumer market. Altho i still feel even then a simple beep every 2 min when in error would be very welcome.

Before you look elsewhere email Limetech. Whilst this is the official forum the community do almost all the support and there is a real chance they wont see or reply to this post (almost certain TBH)

 

I suspect Limetech will consider adding what you need so open a dialog with them

A simple beep when in error would be sufficient for our purpose, but i understand now unraid is more orientated to an consumer market. Altho i still feel even then a simple beep every 2 min when in error would be very welcome.

 

You could add this yourself if you're inclined.

 

There is already, as discussed, an email notify script which fires when an issue with the array is detected.

 

You could modify this script such that in the event of a failure event triggering an email it could also beep the internal speaker until such time as the issue is addressed.

 

A quick test on my desktop shows :

 

echo -en "\007"

 

Will make my internal speaker beep! A very dirty loop issuing that command every second would probably catch someones attention and might be good enough for you?

 

It means hacking it in yourself though - I'm not sure how welcome that would be if this is to be in a commercial environment.

I've been thinking of compiling and create a nagios monitor package (with embedded perl support too).

This could provide all the basics of a monitor system and with other packages an LCD for front end health display.

I'm just waiting for a simpler addon management mechanism.

I'm just waiting for a simpler addon management mechanism.

 

Sorry - very offtopic.

 

Who are you (or we as I agree..) waiting for?

 

It would be very easy to bundle up all the addons made so far as slackware packages. We're then only missing some sort of repository-esque system to allow easy installation and upgrading.

 

http://www.slackpkg.org/index.html

 

Sounds interesting but I don't know anything about it's nuts and bolts. If unraid were based on ubuntu / redhat or something similar with in place advanced package and repository management this would be trivial.

 

 

I'm just waiting for a simpler addon management mechanism.

 

Sorry - very offtopic.

 

Who are you (or we as I agree..) waiting for?

 

It would be very easy to bundle up all the addons made so far as slackware packages. We're then only missing some sort of repository-esque system to allow easy installation and upgrading.

 

A nagios layer would provide a significant monitoring ability to unRAID.

I don't want to write pages of how to install it at the right moment so people do not get false alerts.

It needs to be started when the array is online (or you'll get false alarms).

It needs a mail layer (or LCD display layer).

What I would do is create an addon that is a slackware package and a user drops it in the packages directory. (this is the easy part)

The install script polls the hardware/superblock/mdcmd file and creates the configuration files needed to monitor the system.

Certain information would need to be external, but the bulk of the monitoring can be automated.

This is the harder part and needs to be done at a certain time.

 

Also, the environment would need to be shutdown before the array is shutdown to prevent holding disks open and/or false alarms.

I'm just waiting for a simpler addon management mechanism.

 

Sorry - very offtopic.

 

Who are you (or we as I agree..) waiting for?

 

It would be very easy to bundle up all the addons made so far as slackware packages. We're then only missing some sort of repository-esque system to allow easy installation and upgrading.

 

A nagios layer would provide a significant monitoring ability to unRAID.

I don't want to write pages of how to install it at the right moment so people do not get false alerts.

It needs to be started when the array is online (or you'll get false alarms).

It needs a mail layer (or LCD display layer).

What I would do is create an addon that is a slackware package and a user drops it in the packages directory. (this is the easy part)

The install script polls the hardware/superblock/mdcmd file and creates the configuration files needed to monitor the system.

Certain information would need to be external, but the bulk of the monitoring can be automated.

This is the harder part and needs to be done at a certain time.

 

Also, the environment would need to be shutdown before the array is shutdown to prevent holding disks open and/or false alarms.

 

So it's the promised array API coming in version 5 you're interested in? Rather than better general user script and package handling?

 

I think I mis-understood your post!

So it's the promised array API coming in version 5 you're interested in? Rather than better general user script and package handling?

 

I think I mis-understood your post!

 

Yes..

I mis-posted your understanding. 

 

I know how to make slackware packages.

I know know how to pass runtime configurables to the slackware install package. (to a certain extent).

 

I don't want to write lots of how to information on installing into a go script and/or at what time to install it based on array start.

I think the addons should be simple drop in files.

 

The Nagios environment provides allot of options for monitoring.

However there may be other simpler means. I remember an old tool we used to use called big brother & big sister.

 

Still, any of these requires a mail facility.

Thank u for the responces we purchased a pro license to try the product out, as our test period will be at least 3 - 6 months before we even consider using it comercial there will be some time to fix the poor notifications problems.

Remember to open an ongoing dialog with Limetech. What nearly stopped you evaluating unRAID might be stopping lots more customers and thats feedback any company would want.

Remember to open an ongoing dialog with Limetech. What nearly stopped you evaluating unRAID might be stopping lots more customers and thats feedback any company would want.

 

Just a comment from a rep for a company that bought three fully loaded servers from limetech:

 

The lack of a built in notification when a drive fails is a huge 'minus' for us too.

 

Yes I know I can add this on myself thanks to the community here, and will eventuallly, but I also feel this is a very important missing native feature.