RCC-VE installation
-
Thought I'd crosspost this for anyone not monitoring some of the other threads about this hardware.
Anybody with a new RCC-VE or SG system should probably enable PowerD per this thread if they haven't already!
-
@BlueKobold:
On what did you install? Soldered on board nand, SSD, HDD, or mSATA?
Did you wipe out the pre installed Linux OS from the nand storage?I installed it on a 35 GB mSata-device.
No, I left the CentOS on the NAND-storage. I will probably have no use for it, but you'll never know.
I thought I ordered a dedicated Pfsense-box, but due to some misunderstandings I ended up with this device.@BlueKobold:
What kind of USB stick did you using?
I used a Kingston USB-stick which I loaded with the live installer for 2.1.5
I changed the /boot/loader.conf using my old pfsense box so I could run the installer.
After it was on the SATA I needed to do that again (this time using the USB-stick to boot FreeBSD) on the SATA.
I merely needed to download the latest config.xml from the web (I have an hourly back-up of my pfsense-box) and swap the UTP-cables.@BlueKobold:
Is there really more to it or shouldn't I worry now everything is working?
- upgrading to 2.2.1 perhaps?
On my previous machine and on a test-run the 2.2.1 failed to run properly with the config I have in production.
I did some changes to the config to make use of the 2 extra Ethernet-ports and discovered an orphaned gateway that belonged to an interface that has been deleted several months ago.
Maybe it was the reason the 2.2.1 failed to run.@BlueKobold:
- activating TRIM support for mSATA or SATA SSD if in usage?
Gonzopancho wrote that that isn't needed anymore.
@BlueKobold:- setting up another ether dump buffer for the LAN Ports if this is needed?
Could you elaborate on that?
I have no idea what you mean by that.I do have some problems with the webconfigurator not being available sometimes and ntopng is also a bit flaky.
I'm afraid the CPU is not up to it…. -
FYI, Loading the Development 2.2.2 with a memstick works great to the eMMC, BUT…..
If you pull the power and reboot several times, you'll see the file system get more and more corrupted until it doesn't boot anymore.
Perhaps a case where you should be putting nanobsd on this?
-
I don't think nano is all that and a bag of chips. It has it's place.
I'm not sure why you love pulling the power on your box, either.
-
In the field, at a customer site, you don't always have a nice environment where you don't have power outages. Even with UPS systems, you still have the real world situation of being able to survive a power outage.
I have a couple of hundred Alix boxes (nanobsd) in the field, a few dozen APU's and even some FW-7541's
My concern isn't with trying to defend my new toys, but to be able to support my customers. The issue I have posted is a real problem for this box and makes it unusable outside of the lab unless the issue is resolved. With this issue, in this configuration, it is untrustworthy. If that is not important to you, then good for you.
-
With this issue, in this configuration, it is untrustworthy. If that is not important to you, then good for you.
I love it when people attempt to ascribe things to me that I did not say.
When I was the CTO of Wayport (now AT&T WiFi Services) I had 25,000 systems in the field running linux, and that didn't include the purpose-built managed Ethernet switches (some with 10/100 Mbps PHYs, some with VDSL PHYs) that also ran linux. (FreeBSD wasn't an option then.) IIRC, there were over 100K linux boxes under management in the field, never mind the 'back end'.
So… I am aware of the issue(s), and I know how to address them.
You don't care as much about nano as you do a stable filesystem. The thing in the 'nano' images that gives you that today is the
fact that all the disk-resident parts of the system are mounted read-only. Future direction is a much more stable filesystem. -
I asked a simple question, if nanobsd wouldn't fit better due to the OS corruption caused by power loss. Perhaps the way I asked it, you took it a different way. It would seem instead of addressing the actual issue, you wanted to dance around your own agenda.
You then chose to be accusatory and make light of "pulling the power" on the system as if I liked to do so. It is called testing. It failed. I added that information so it might help someone else down the road as well as seeing if there is something I might be missing. The question at this time is how this gets addressed.
I would think the whole idea of the forum is to share information so we may all achieve useable systems. None of your answers to my original comments are "helpful" in any way and I'm not really sure why you seem to be taking those comments as some personal attack which is definitely not my intent.
-
Maybe I missed it in all the banter here, but can someone confirm:
a) Is it possible to get a nanoBSD install onto the internal flash of these boxes?b) If so, is there a published procedure to do it? Is it complicated or easy?
The sun goes down, the battery is old and goes flat, the guy in some remote office just likes pulling the plug,… stuff happens. I like the nanoBSD because I have never yet had a problem as a result of unexpected power loss.
-
the guy in some remote office just likes pulling the plug>>
Ain't that the way it usually works. The other day we had a contractor pull the plug on a customers phone system so he could plug in his drill. You never know.If I can figure out some way to do it, I'll post.
-
a) Is it possible to get a nanoBSD install onto the internal flash of these boxes?
Would this post from jimp eventually answer your question? eMMC on SG units
b) If so, is there a published procedure to do it? Is it complicated or easy?
I really think that only very experienced users, the development team or the guys from
netgate or pfSense shop and last but not least with hands on those boxes would be able
to draw a moderate way that is not harming the nand storage and is also running smooth.I like the nanoBSD because I have never yet had a problem as a result of unexpected power loss.
For sure this will be related to the circumstance that the nanoBSD image is mounted "read only"
and there for it would not be something going wrong with this image in my opinion.