Suggest: view current drive temp in FreeNAS UI

[diskdiddler]

Hi All,

I'd love to be able to view the current disk temp in the FreeNAS UI, for the life of me, I can't figure out how to do this. I know the alerts handle hot disks but it'd be nice to be able to visibly see it in one of the menus the disks are already listed in

Thanks

 

[Z300M]

Hi All,

I'd love to be able to view the current disk temp in the FreeNAS UI, for the life of me, I can't figure out how to do this. I know the alerts handle hot disks but it'd be nice to be able to visibly see it in one of the menus the disks are already listed in

Thanks

Add my vote for this.

 

[cyberjock]

This has been discussed at least twenty times in the last two years.

The reality is that it's a "nice to have" it really doesn't add as much value as you think. If you've done your job as a server admin and setup SMART to monitor your temps it will email you if your hard drives get too hot (with "too hot" being over whatever threshold you want to set).

As a result, there's not much reason to add the feature and it's basically been on the "back burner" for a while. If you want to add it I'm sure iX would be happy to integreat your pull request. ;)

 

[diskdiddler]

What do you mean it doesn't add value? If you set up your machine fresh and need to see the figures quickly before finalising the enclosure, packing it into a noiseless space, etc - it's very useful.
If it's been discussed at least twenty times in the last two years, there's clearly demand for this.

 

[cyberjock]

What I mean when I say "it doesn't add value" is that it doesn't really add much more value than simply warning you if temps are above your pre-determined threshold.

For example, I get an email at 37C. Do I care if it's 30C or 20C? Not really. It's not overheating so I don't care. And let's be honest with ourselves. What's more likely to catch an overheating drive? The admin that logs in just more frequently than never or a monitoring program like SMART that is checked at regular intervals?

If you *do* feel the urge to see the actual temp, smartctl will let you get the temp. But the reality is that as long as it's not "too hot" for whatever definition of "too hot" you want to use is, then the disk is okay.

Sure, there's a bunch of edge cases, but spending hours adding a feature for that occasional edge case isn't exactly maximizing limited developer resources.

And you'd be correct that "there's demand" for it. But just because there is demand doesn't mean there is an urgent need either.

If you've done a proper setup of your server you should never *need* to monitor it. If you *want* to monitor it that's one thing. But the need isn't there because FreeNAS is designed to email you if anything goes wrong. iXsystems is trying to solve major problems with very limited developer resources. So "needs" far outweigh "wants". The ticket does exist in the system and has for quite some time. It will hopefully be added someday. When extra developer resources exist for "wants" then it might be added. Or if some free-lancer wants to add the feature I'm sure iX would appreciate the support. But, until then the needs that FreeNAS has outweighs the wants. ;)

Just as a thought, do you know how many times I've looked at the reporting tab on my own server this year? Never. Not once. Do you know why? Because my FreeNAS box is properly setup to email me in the event there is a problem. If they added SMART temps they are something I'd look at when I upgraded my FreeNAS box, thought "how cool" and then promptly never looked at again.

 

[diskdiddler]

and if you change internet provider and forget your SMTP server has changed?
There's a plethora of reasons why clicking and spotting something like that while hunting through the GUI would be useful to users, I hardly think it's an even remotely unreasonable request, the system is already capable of not only pulling the data but emailing it to the administrator. The system pulls serial numbers and presents them on the view disks page.
There's all kinds of information on "view disks" yet temperature is omitted, sorry but it makes no sense to me at all.

 

[cyberjock]

Thank you! You just validated my arguments.

FreeNAS used to send nightly emails. But a bunch of whiners complained and now they are gone. I argued against getting rid of the nightly "everything is okay" emails and lost. So now I have a custom script that sends me night emails. I now know that if I stop getting them suddenly then something is wrong.

But, if you fail to change your smtp then you are broken. It's also self-inflicted and many would argue "you can't fix stupid". Also if you look around all the other monitoring systems use smtp too. So if you are making that argument you've already failed to properly maintain your other monitoring systems and you deserve whatever pain you get as a result of your lack of planning. Because there's flat out no excuse for not maintaining smtp settings in your machines that do their own monitoring. When I've worked in IT we kept a list of all of the places that used smtp and when we had to change account info we had to go through the checklist and update all of them.

I don't know about you but if you are logging into boxes because you are bored and just "monitoring" you have probably done a lot of things wrong elsewhere. 99% of the admins I know are far too busy to be dealing with logging into boxes "because they can". This is why monitoring programs are worth their weight in gold. They monitor so you don't have to. There's nothing that you would do in a webgui that a good monitoring program can't do far far far better than you'd ever be able to do. I'm pretty sure you aren't logged into the server 24x7 monitoring non-stop, yet you argue that the one login you want to make matters. Well, if you want that one login to matter the CLI is waiting for you.

You aren't going to convince me if your side and I'm pretty sure I haven't convinced you of my side. The bottom line is the devs aren't adding it right now and may not for quite some time. If this is a show-stopper there are plenty of other OSes you are welcome to use. FreeNAS is built for simplicity and effectiveness and currently looking at hard drive temps isn't one of those.

 

[diskdiddler]

Short sighted. Sorry, not convinced no. Close thread - you can do all the same replies to the next person who asks for it, because it's really not a big deal.

 

[enemy85]

Just type (when u need) "smartctl -a /dev/adaXX | grep Temperature" from Shell and u'll solve the problem...

 

[no_connection]

Some of us want to log into the box one in a while and check if everything is behaving right.

If the temps are up two degrees from what thy usually are then it's time to check air filter. Or external cooling.

Using smartctl is a minor inconvenience so I rather see the devs do something else. But as they are redesigning GUI they might as well add some features while there at it.

 

[cyberjock]

Some of us want to log into the box one in a while and check if everything is behaving right.

If the temps are up two degrees from what thy usually are then it's time to check air filter. Or external cooling.

Using smartctl is a minor inconvenience so I rather see the devs do something else. But as they are redesigning GUI they might as well add some features while there at it.

You realize most people's hard drives fluctuate by more than 2 degrees just by running a scrub or otherwise being kept busy? If you are like me and your server isn't stored in a *strictly* temperature controlled room you can see more fluctuation than 2 degrees just because of the time of day?

So no, 2 degrees isn't "time to check the air filter". ;)

 

[no_connection]

I sad "from what they usually are". And no, they usually don't deviate more than 2 degrees.
You don't need temperature controlled for that. For you it might be different but for me it's very stable.

I check temperature right now and it's right where it usually is. But I almost couldn't, smartctl lists too many SMART tests and I can't scroll in webshell.
So soon I need to SSH into the NAS to check which is just stupid as I have to enable that.

So it really needs to be added. At the very least a SMART output screen, as I can't view all per above.
You can keep arguing against it but it just went from nice to have to needed.

 

[cyberjock]

I didn't argue against it. I'm simply saying its nothing more than a nice to have and backed it up with cold hard facts for why that is the case. As for your filter, *even* if you were correct, it heating up 2C is not reason to clean the air filter. Temps will take months to creep up to any kind of temperature you should be worried out in normal conditions, so set your temperature for the emails to a suitably low level and forget about it. Problem solved! There, I just fixed your "dillema".

It really doesn't matter if 1000 people show up here arguing for it or not, the devs have more important things to do so its been given a low priority. I'm not the decision maker on what is or isn't included in updates to FreeNAS, and I'm glad I'm not. That guy has a shitty job that will be thankless and never make everyone happy.

 

[danb35]

But I almost couldn't, smartctl lists too many SMART tests and I can't scroll in webshell.
So soon I need to SSH into the NAS to check which is just stupid as I have to enable that.

I agree that it would be nice to monitor this on the "reporting" tab, but this seems really over-the-top. Why not just enable ssh? It has to be easier to ssh into your server than to log into the web GUI, then fire up the shell, and then not be able to scroll or whatever.

Or just use the "more" command, like "smartctl -a /dev/ada0 | more". No need to scroll then.

Or use grep: "smartctl -a /dev/ada0 | grep Temperature".

 

[fracai]

I put in a ticket a while back asking for the ability to upload the data from the reporting tab to Librato. It's not a high priority and who knows if it'll even ever be implemented, so I took matters into my own hands. I developed a python script that reads the RRD files and uploads the data. I then started parsing proc and smartctl to add CPU and drive temperatures.

The result is a single page (Librato only offers a single dashboard for free accounts) with a lot of pretty graphs. As a bonus, if the script stops reporting, Librato can send you an email. Maybe the script got hung up (bug reports and pull requests will be appreciate), but maybe there's a server issue.

Anyway, I'll smooth off a few corners and post the script. It needs some additional python modules, including one that must be installed from source rather than pip or ports, so it'll need to be run from a jail and I'll need to put together something on installing the script and dependencies. I'll try to get everything done tonight, but I may not get it finished until the weekend.

 

[Knowltey]

What do you mean it doesn't add value? If you set up your machine fresh and need to see the figures quickly before finalising the enclosure, packing it into a noiseless space, etc - it's very useful.

GUI -> Shell ->
smartctl -a /dev/ada0 | grep empe
smartctl -a /dev/ada1 | grep empe

so on and so forth

 

[no_connection]

For us that used FreeNAS before iX rewrote it it's a bit hard to understand why they didn't include the level of information that was available in GUI.
Temperature next to every disk, page with SMART for every device that supported it, list of open files, sorted log that you could download, including disk temperature.
It's not like they had to reinvent the wheel to get it. Being very easy to get the data itself, the hard part would be to implement it into the new GUI, which is why it's a good time to do it now when they are redesigning it.

An advanced information tab with a bunch of badly formatted data would suffice as long as it's an easy way to view SMART and open files. If it's ugly then hide it with an advanced checkbox.

 

[cyberjock]

Actually, they did have to reinvent the wheel. The whole problem was that the old code wasn't really written with the expectation of being "forward" compatible and "forward" thinking. It had major limitations and problems internally and so a rewrite without reusing any code was necessary.

The thing to keep in mind is that just querying more than 10 disks can take upwards of a minute or two. As TrueNAS is scalable to hundreds of disks can you imagine a WebGUI that either is only updated occasionally (say once an hour so is almost always out of date) or simply locks up the WebGUI for 10 minutes while it queries your dozens of disk?

There are technological limitations to this so we can't just throw this out there and let it work in some 1/2 ass state.

 

[fracai]

To be fair, the GUI thread shouldn't be collecting the stats. Let that happen in the background.
</pedant>

 

[Whattteva]

I'd be fine with most recent cached temperature instead of current.