Here is why NAS functionality on pfsense can make a hell lot of sense.
-
Well, maybe I'm gonna use older hardware that I already have lying around? I don't know, maybe I won't have to buy anything.
I'd be interested in seeing how much power the vm will draw when 3 clients copy files from it at the same time with full gbit speeds each. And I'd also like to see how much CPU power that would require.
-
why would the power draw go up??? Because the cpu works a bit more??? Not going to do anything significant.. I would have to put my killawatt meter on it if you want to see if it changed by a watt or so..
Here I just moved 10GB.. You can see the cpu of the vm spike up a bit.. still not sucking up what its been given even.. Here is my power graph from yesterday… Was watching quite a few videos in the evening.. Don't see any spike in the power draw..
-
The other way to look at this "problem" is that it has been already solved by others.
Look at other open source alternatives, DD-WRT, OpenWRT, etc.
They have Samba built-in for those who desire to go down that path.pfSense has built itself on a different philosophy and IMHO has been very successful because of it.
As mentioned, in this and many other threads on this site, the consensus of opinion is:
The potential risks don't justify the possible rewards.
As also said, (I paraphrase) "The horse is dead Jim…."
-
@johnpoz
I find it very hard to believe that there wouldn't be a significant difference in CPU usage and power consumption when transfering data to clients with a total of 3Gbit/s.
After all 10GbE requires "i7 or Xeon CPUs", according to the forums.I mean I don't know what kind of a CPU would be required for 3Gbit/s, but let's stay at 10GbE for a second.
Say I'd like to put a 10GbE card in that router. Now wouldn't both, the host machine and the vm need an "i7 or Xeon CPU"? (That's a serious question because I really don't know.)
I mean I would guess that the data would have to be moved from the vm to the host and then from the host to the client. And that first transfer from the vm to the host would waste a lot of resources. (As I said correct me if I'm wrong.)But okay sure I won't need 10GbE. So let's say the router would need at least the power of an "i3 CPU" for my 3Gbit/s. According to my theory the NAS would need it roughly two times the power then. So an i3 wouldn't be enough anymore and i5s or i7s are significantly more expensive.
@divsys
Well, I was told by multiple people that pfsense is basically by far the "best" router OS.
And about the potential risks, I already made my comments in the first post.
Besides the fact that there would be no changes in security for people not using the feature, I doubt that there is a realistic chance of something like this creating new vulnerabilities; unless you were to implement it very poorly.
And as a consumer, with no valuable data, it wouldn't even really care if my data got into the hands of a random hacker. -
@johnpoz
I find it very hard to believe that there wouldn't be a significant difference in CPU usage and power consumption when transfering data to clients with a total of 3Gbit/s.
After all 10GbE requires "i7 or Xeon CPUs", according to the forums.I grabbed this piece of your answer because I believe it underlies a little bit of spec's "cherry-picking" to prove your case (it was probably unintentional on your part).
When you/I/others 'see' a comment like "you need i7 for 10GbE" in a pfSense thread, remember what pfSense is about - routing packets (or more precisely "filtering" them the pf in pfSense).
That implies that pfSense is going to examine,analyze,and direct ALL the traffic it sees on any of it's interfaces. Everyone of those operations takes CPU horsepower the faster the packets the more horsepower you need, that's the way pfSense was designed.When you talk about a Samba/NAS/File share device, it's job is much simpler from a "packets" POV - respond to the packets that have already been cleared for it (by pfSense and others).
There's all sorts of stuff it may do about rights/permissions/etc, but as far as CPU Power required to move data across a NIC, it's pretty minimal compared to what you're asking pfSense to do. The actual transfer of data is a matter of making sure (broadly speaking) the NIC buffers don't get empty. That takes some more HP as speeds increase, but no where near as much as for examining every packet you see.The net result is that I'm not surprised at all by johnpoz's results, they match every VM setup I've worked with. They also explain why VM installations are becoming the norm more than less so. Most modern hardware has tons of excess capacity and VM makes it easy to leverage that into more capabilities.
…Well, I was told by multiple people that pfsense is basically by far the "best" router OS.
And about the potential risks, I already made my comments in the first post.I would argue that the very reason it's by far the "best" router OS (I do agree largely BTW) is that they recognized very early on that being the "best" doesn't mean being "everything"
Sorry, it's still dead Jim :)
-
@NopIt
I'm at a loss at to your responses. I don't want a flame war but I won't run away from one as well.The way I see it you want it both ways and will jump to either side when it's convenient for you. You want a system that gives the ability to run multiple OS but want to save on equipment cost and power cost. What modern day equipment offers a single core CPU or SCO? None come to my mind. Even today's current Raspberry Pi 2 has multiple cores. You want a system that has decent response time AND while answering ALL the demands of a router and a NAS. Yes I know what a NAS is as well as a SAN, DASD, etc. further, I dare say most commercial NAS systems don't reach true gigabit speeds. So shut off that smoke and mirrors.
You then jump to the fact you'll just use one of your older equipment. Great! Exactly how is that going to save $? I mean let's talk about energy efficiency. Today's equipment is way better than something of 5-10 years ago. Also, don't forget to get that 486 or single core Pentium fully stocked on RAM so it can do all that you ask of it with decent response times. And, don't forget to code tightly for CPU optimization, you don't want ANY memory leaks that will shut down/crash the ENTIRE system!
Prior to me bringing up the software for free you NEVER mentioned paying for it. Show where you did prior to my mention. Again, both ways.
You make mention you have a family, work, and a life and you don't have time to learn, etc. Oh please get over yourself. How does that make you any different from someone else with their lives? You think someone else does? Rhetorical questions. If you want it you'll invest the BST (Blood, Sweat, Trauma) to coding what you desire and you can offer it for free or charge for your efforts. But remember, as of this post date no one has done it, and if they did you wouldn't be asking for it!
CMB, responded with VM is best so that all the services of each OS really doesn't work well when combined together, doktornotor is right this is a complete was of DB space, johnpoz gave great examples of why this horse has been beaten to death along with you really not knowing about VM and how his VM is doing all that it does with modest power usage (IMO). (BTW your cheap shot of "that you straight up don't know enough about all this." Opinions vary, but run on with your bad self to johnpoz - Good luck.), and Jailer great pic of advise.
Bottom line - there are some REALLY REALY smart people here giving their true opinion(s) and the reasons why, but you just keep your head in the sand.
Stay there, but for me I'm moving on to things that make more sense than your flip flopping, or ignoring solid advice.
-
@pfSense4ME
I said I might use some older hardware that I already have.
There are a few calculations to do before I can say what would be the smartest move for me.
I mean let's say my older hardware would cost me 150€/year and the newer hardware would only cost me 100€/year, but 400€ to buy it.
Now if I would wait a years and the hardware would then only cost 300€, I could upgrade my system then and save 50 bucks withing the first year.
I'm just saying it's not that easy.Yes, I DO think that someone else will have the time. It would cost me month, if not years of work to do something like this.
I will definitely try to find a developer who actually knows what he is doing and pay him to get it done.@divsys
Thanks for shining some light on all this.
That may change everything for me.Okay, I've got a Intel Pentium G3258 lying around that I bought form a fried for very cheap a while ago. I just looked it up and it lacks VT-d (Intel Virtualization Technology for Directed I/O). And that even though it is from 2013 or so..
That's kind of what I meant earlier. Now how would that affect my FreeNAS-in-a-vm-experience? What are my disadvantages of not having VT-d?
And in general what happens if I want to swap a hard drive in my NAS? Would I have to reconfigure the whole vm for that? Or could it just handle it automatically?And would such a CPU be enough? After all it's just a dual core part and I'm not sure if I'd want to overclock it (fan noise).
-
It's a security risk to store your files on the same machine as the firewall. If your firewall machine gets hacked the hacker could gain access to your data.
When I posts like that I was shocked. I personally live in a thing called democracy. If I want to buy and eat something that makes me fat, then I can do so.
You're free to make your own crappy Firewall+NAS, but most people here won't help you harm yourself.
I've wanted a NAS for over 15 years, but I refuse to make one until I can make it correctly. I'd rather do without than half-ass it. Do it correctly or don't do it at all.
-
I've wanted a NAS for over 15 years, but I refuse to make one until I can make it correctly. I'd rather do without than half-ass it. Do it correctly or don't do it at all.
For some of the current definitions of "correct" you can look at http://www.freenas.org/ or http://www.nas4free.org/.
-
It clearly supports vt-x, why do you need directed i/o vt-d –Unless you need to give 1 vm specific access to some hardware, it is not needed.. Sure an the hell not need to run a nas and your router on the same host that is for sure..
Talk about cherry picking info.. Vt-d is going to be included in their HIGH END cpus.. Not some msrp $72 cheap budget cpu.. It supports most none of the advanced features
http://ark.intel.com/products/82723/Intel-Pentium-Processor-G3258-3M-Cache-3_20-GHzThat chip released Q2 2014, not 2013 btw..
Vt-d didn't even come on the table until end of 2008.. Pick any HIGH END cpu after that period and it will most likely support Vt-d.. My ford focus doesn't have a turbo charger either, so what I can not drive it?
VT-x has been around atleast 10 years.. And is included in almost all current cpus, yes even that budget chip you pointed out..
My system doesn't support aes-ni either, but guess what it still does openvpn just fine.. You don't always need a freaking Ferrari to drive to and from work..
-
I've wanted a NAS for over 15 years, but I refuse to make one until I can make it correctly. I'd rather do without than half-ass it. Do it correctly or don't do it at all.
For some of the current definitions of "correct" you can look at http://www.freenas.org/ or http://www.nas4free.org/.
My definition of "correct" is the physical hardware. I figure I need at least $2k to get started. I won't go for anything less than 1TiB of logical, back by all SSDs of several different brands, Xeon, 10Gb NIC+switch, and 64GiB of DDR4. The bigger issue is finding some good hot-swap hardware(bays). Most stuff that I can find on NewEgg has people complaining about cheap parts and the plugs not aligning, plugs breaking, general connection issues resulting in a drive suddenly disconnecting.
My alternative is to just get something from iXSystems.
-
"10Gb NIC+switch"
There you just blew your 2k$ budget ;) 10Gb switches are not really home/lab budgeted yet.. Atleast not that I have seen.
I was looking at the new supermicro http://www.wiredzone.com/supermicro-servers-compact-embedded-processor-sys-5028d-tn4t-10024470 that you can get for $1200 without anything, but does have dual 10G nic via soc and 2 more gig nics.. The problem is the switch to connect it at 10Gb ;) Will do up to 128GB ddr4, would be a screaming vm host.. Once you put some memory in it and some disks your pushing the 2k budget.. But those 10G nics would be nice future proofing for when the 10Ge switches get to be more reasonable.
-
@Harvy66
"You're free to make your own crappy Firewall+NAS, but most people here won't help you harm yourself."
Can you even read? HOW WOULD I POSSIBLY HARM MYSELF?
I said it a million times now. I'm a private user, no one would benefit from explicitly hacking me. The data on the NAS will be stored encrypted just because I can. I wouldn't even really care if a hacker would get my data.
But all that doesn't even matter because running a NAS (in a vm) on pfsense does not create vulnerabilities in pfsense, unless pfsense by itself is a poorly written piece of crap, which I highly doubt.
But if you are so certain that security would be affected that drastically, proof it.@johnpoz
Calm down, please.
I'mwell aware that it supports vt-x and I never said I need vt-d. I was just asking nicely what the disadvantages of not having it would be in my case.
"Q2 2014" - I don't want to impute nitpicking to you, but that's really not relevant. Besides I said "2013 or so"; I was just estimating.
"Unless you need to give 1 vm specific access to some hardware"
Well yeah, how about the hard drives that go into the NAS? Will I have access to S.M.A.R.T and could I create a file system on the drives from the NAS OS without vt-d? And what about pcie raid controllers? -
"I was just asking nicely what the disadvantages of not having it would be in my case."
None - unless you wanted to directly connect some hardware to a vm..
Well it would depend on how you connect them, I don't have vt-d and I have access to the smart info because I raw map them to the vm.. Would also depend on your hypervisor I would also assume on if it allow for such raw mapping.
Are you talking to the vm OS itself, like esxi? Or the VM? Both can do it - esxi added function in like 5.1 I think, and the disks I raw map to my nas os vm, can see it as well.
In my nas vm (2k12r2) I run some software from stablebit that does my pooling for me, not really a fan of drive spaces for simple home use pooling of disks, and also user their scanner software that watches smart, keeps an eye on the filesystem and disk and sends me an alert if something seems odd, out of wack, etc..
-
Personal opinion, but I really like my firewall to do one thing, one thing only. Much easier to verify correctness, less to loose if something goes bad.
-
My definition of "correct" is the physical hardware. I figure I need at least $2k to get started. I won't go for anything less than 1TiB of logical, back by all SSDs of several different brands, Xeon, 10Gb NIC+switch, and 64GiB of DDR4. The bigger issue is finding some good hot-swap hardware(bays)
From my POV you've stepped from the "Build my own NAS" to the "Build my own Server" maybe a change of perspective is in order…...
-
Many consumer routers come with a USB port to simulate a NAS type of storage, but many consumer routers also have security problems related to this sort of technology. I believe the developers of PFSense could do it correctly, but it really feels like the wrong type of feature to implement on the PFSense platform.
PFSense is designed to be an expandable/modular driven firewall solution for protecting 1 or many networks. Developing "frills" isn't a way forward for a product with a strong focus on security and stability.
If an "All in one" solution is something you would prefer, then I would suggest a typical off the shelf product and flash it with dd-wrt/tomato or a variant if you feel the need.
I think it's noteworthy that you hold PFSense in such high regard and wish to use it as your "All in one" platform, but demanding that a product team implement a feature that hasn't gained traction for obvious reasons is the wrong way to solve your problem.
-
I don't know who or what to believe anymore.
I just read this forum post on the FreeNAS forums:FreeNAS is awesome. FreeNAS can and will run as a VM. That does not make it a good idea.
-
FreeNAS is designed to run on bare metal, without any clever storage systems (UNIX/VMFS filesystem layers, RAID card caches, etc!) getting in the way. Think about this: ZFS is designed to implement the functionality of a RAID controller. However, its cache is your system's RAM, and its processor is your system's CPU, both of which are probably a lot larger and faster than your hardware RAID controller's cache!
-
Without direct access to the hard drives, FreeNAS lacks the ability to read SMART data and identify other developing problems or storage failures.
-
A lot of the power of FreeNAS comes from ZFS. Passing a single virtual disk to ZFS to be shared out via FreeNAS is relatively safe, except that ZFS will only be able to detect and not actually correct any errors that are found, even if there is redundancy in the underlying storage.
-
There is a great temptation to create multiple virtual disks on top of nonredundant datastores in order to gain "MOAR SPACE!!!". This is dangerous. Some specific issues to concern yourself with: The data is unretrievable without the hypervisor software, the hypervisor might be reordering data on the way out (which makes the pool at least temporarily inconsistent), and the hypervisor almost certainly handles device failures non-gracefully, resulting in problems from locked up VM to unbootable VM, plus interesting challenges once you've replaced the failed device.
-
Passing your hard disks to ZFS as RDM to gain the benefits of ZFS and virtualization seems like it would make sense, except that the actual experiences of FreeNAS users is that this works great, right up until something bad happens, at which point usually more wrong things happen, and it becomes a nightmare scenario to work out what has happened with RDM, and in many instances, users have lost their pool. VMware does not support using RDM in this manner, and relying on hacking up your VM config file to force it to happen is dangerous and risky.
-
FreeNAS with hardware PCI passthrough of the storage controller (Intel VT-d) is a smart idea, as it actually addresses the three points above. However, PCI passthrough on most consumer and prosumer grade motherboards is unlikely to work reliably. VT-d for your storage controller is dangerous and risky to your pool. A few server manufacturers seem to have a handle on making this work correctly, but do NOT assume that your non-server-grade board will reliably support this (even if it appears to).
-
Virtualization tempts people to under-resource a FreeNAS instance. FreeNAS can, and will, use as much RAM as you throw at it, for example. Making a 4GB FreeNAS VM may leave you 12GB for other VM's, but is placing your FreeNAS at a dangerously low amount of RAM. 8GB is the floor, the minimum.
-
The vast majority of wannabe-virtualizers seem to want to run FreeNAS in order to provide additional reliable VM storage. Great idea, except that virtualization software typically wants its datastores to all be available prior to powering on VM's, which creates a bootstrap paradox. Put simply, this doesn't work, at least not without lots of manual intervention, timeouts during rebooting, and other headaches. (2013 note, ESXi 5.5 may offer a way around this.)
I'm pretty sure I'm forgetting a few. But the conclusion is this: it's perfectly fine to experiment with FreeNAS in a VM. However, if you run it in production, put your valuable data on it, and then something bad happens, and you absolutely positively must get your data back, there probably won't be a lot of help available from the forum. We've seen it happen again, and again, and again. Sigh.
So is it a bad idea or not? I mean that post does clearly imply that it is. Now the question is how biased each party is..
-
-
More a subject to be hashed out on the freenas forums. Why are you here looking for FreeNAS expertise?
In case you haven't figured it out, it doesn't appear that anybody here is interested.
If you want pfSense and FreeNAS on the same hardware, virtualize (at your own risk.)
Your aforementioned old hardware will almost certainly lack VT-d support.
-
I've not delved too deeply into the current state of the art of FreeNAS so you may take my next comments with some grain of salt.
That said, I have noted there seems to be a current "battle" between FreeNAS and Nas4Free (a recent fork of FreeNAS) on a whole range of issues, some of which you've touched upon.
As far as the include a "NAS in pfSense" debate, everything you've touched on is evidence in my mind NOT to include a NAS in pfSense.
Perhaps VM is the way to go, perhaps not, but it's pretty obvious to me from your quote that a NAS has really different issues of concern than a firewall.
No point in trying to shoehorn them together.As Derelict mentioned, I think you've moved off of this being a pfSense issue.
I'd suggest a little research into the current NAS distro issues and where you want to go from here.
