[R7800] Will no longer save settings after reboot.

[L&LD]

@HELLO_wORLD it may seem counter-intuitive, but USB 2.0 isn't 'too slow' for running scripts or even a swap file (at least for Asus/RMerlin powered routers). Yes, it is slower than USB 3.0, but only when setting up or updating scripts, not when running them.

 

[Droidrat]

Run point 3 + 4 from here:
https://www.snbforums.com/threads/kamoj-add-on-v5-for-netgear-r7800-x4s-and-r9000-x10.60590/

Thank you.

FYI. I'm using Voxel V1.0.2.75.2SF
The response to "artmtd -r protime | strings" was "Protime:" ... i.e. "really old"
The response to "date -d $(nvram get config_timestamp) -D '%s' "+%F %T"" was "2020-03-31 07:14:19"
At one point, I used Traffic Meter.

While it is possible that Traffic Meter "exhausted" my NVRAM, I find it strange that the same problem is being experienced by a number of people at the same time.

I have attached a snippet showing my memory usage (via the Kamoj addon; thank you @kamoj ). It shows NVRAM in use, and gives a correct total for the total NVRAM. If the chip was toasted, would it still display like this?

The installation of the addon necessitated a "reboot" command. No settings were lost.

 

[verbage]

So, I have experimented, and I have confusing results...

Very interesting, indeed, right before returning my original R7800 with failed NVRAM in the RMA swap, I decided to try it out one last time. And lo and behold, the darned thing seemed to remember its settings--even after completely shutting it down, unplugging it, etc.!!! So after sitting unplugged and unused for several days, it seemed to recover...

I had flashed it down to the latest official firmware at that point (i.e. v1.0.2.68) for return to Netgear. But with that same .68 firmware in the previous days while it was still in continuous operation, it had not preserved the settings. So it does not seem to be related to firmware version.

I do not understand why it started functioning again. The only logical explanation I can think of is that the issue might be heat-related... When it originally lost the settings, it had been on and running continuously. Over the next few days, I kept it plugged in and running as I worked out the problem, and then waited for the RMA swap to arrive. So it was always on, and thus, always warm/hot, during that time. And if I did a power cycle, it would lose its settings. Only after letting it sit around unplugged and unused for several days did it seem to recover. For that reason, I speculate that the issue may be heat-related.

The saga continues because with the brand new RMA swap R7800 I received, I had it up and running with Voxel's latest v1.0.2.75.2SF firmware, and it was working great. Even on full power cycles with full unplugging, it came right back up as expected with a preserved configuration. But then after just two days (on Thursday afternoon), it failed in exactly the same way as my original R7800. I.e. out-of-the-blue*, it was reset to factory settings, and then would not remember the configuration on a power cycle, nor was it able to ingest any of the saved configuration files I had made over the last two days.

* I wrote "out-of-the-blue" in the last paragraph, but it was not completely random. We suffered an electrical brownout at that moment, and I heard the UPS feeding the router make a single beep before it recovered. The voltage sagged enough that the router lost power, and that is what seemed to have caused the problem. This indicates that the UPS battery was probably failing so I swapped in a whole new UPS unit. I suspect this is also what happened about two weeks ago when I was at work, and got the emergency message from home that "the Internet was not working."

Anyway, on Thursday afternoon after the failure with the new RMA swap R7800, I had no option but to pull it and substitute in a backup router to maintain peace with the family. And I left the R7800 unplugged sitting in the corner, and feeling totally dejected that it has lasted only two days. Then, yesterday evening, after it sat unplugged for two days, I figured I would try it again...

I did not boot it directly to check out how things were. Instead, I had a hankering to try out OpenWRT on it so I immediately flashed that, and did a factory reset after that. Then I booted it up, started playing with it, did full power cycles including full unplugs, and was stunned to see that it was preserving settings at that point!!! Even though it had not been doing it just two days earlier!!! What was the difference? Nominally nothing except that it had sat unplugged in the corner for two days. So this somewhat supports the idea of a heat-related problem.

If it is a heat-related problem, it should affect any firmware--official, Voxel's, OpenWRT, or whatever, right? So I am doing a test now, and have left it with OpenWRT. If it fails with OpenWRT, well, for sure, it is a hardware problem.

The other difference, as mentioned above, is that I have swapped in a new UPS. But gosh darn it, it seems strange to me that the device would be so susceptible to power fluctuations. And after the brownout, it clearly could not maintain its settings nor read in old config files. Anyway, I will keep testing it with OpenWRT to see what happens.

One slick thing about OpenWRT is that all the config files are simple text files. So I was able to grab the /etc/config/dhcp file from another OpenWRT install, in which I had previously spent the time to change all the machine names to something recognizable. I just scp'ed it over to R7800 which I had just flashed with OpenWRT, rebooted, and had all my machine names set up!!! This saved me a huge amount of time!!! It would be cool to be able to do this with Voxel's firmware, too, because trying to restore settings via the web GUI does not seem to work**.

** Actually, it just occurred to me that if the issue is heat-related, maybe after waiting till the R7800 cools down and rests, the restoration of a configuration via the web GUI would work, too?

 

[OllieNZ]

Does there seem to be a build date or use type this started from? Mine's at least 2 years old (purchased in 2018) I don't play with it beyond having installed Voxel's FW and setting my own SSID's. I don't use traffic management. It is heavily loaded from a client perspective with multiple smart phone, pcs, smart tvs, consoles and a NAS.
I suppose what I'm asking is what's the likely hood of seeing a failure of my router?

 

[verbage]

I suppose what I'm asking is what's the likely hood of seeing a failure of my router?

Good question--does this equally affect all R7800s from any time/date range of production? Or does it only affect a particular batch of R7800s? In the latter case, this could be explained, for example, by a batch of slightly less robust NVRAM chips that made it into a finite time/date range of production by Netgear.

Two things to keep in mind--this does not affect just the R7800... If you do a search on "netgear router loses settings" (https://www.google.com/search?q=netgear+router+loses+settings), just in the first page of results, you'll see lost settings mentioned in relation to a whole bunch of Nighthawk router models including the R7500v2, R6400, R7000, R7800, R7900P, R800P, and R9000 for good measure!!! Maybe they all use the same NVRAM chip (easy enough to figure out), and thus, maybe all can be pinned to a finite time/date range of production, which might indicate a batch of slightly less robust NVRAM chips?

Keep in mind that even though all those different models are mentioned with respect to lost settings, without reading any further, the specific symptoms and manifestation of the issue might be different from that which is discussed in this thread. For example, in some cases, it seems like folks were able to reload old configuration files, and quickly get back to business. But at least in my experience, and that of the others who have mentioned the issue in this thread, we were/are unable to reload old configuration files, which makes the problem super-annoying because it means having to painstakingly retype in one's whole configuration...

The other thing is that there seem to be a lot of hints that this issue is often related to power outages or failures--the NVRAM chips and supporting circuitry seem to be sensitive to this. For sure, this cannot be pointed to in all cases, but unless one is specifically next to the router when something like this happens, it might be hard to know. As I mentioned, with my new RMA swap R7800 that failed on Thursday, I was in the adjacent room when we had the power brownout, and I actually heard the UPS give one beep, and then, the R7800 went kaput. I suspect this also might have been what happened to my original R7800 a few weeks back, too, but I was simply not there to be able to confirm it.

So what does this mean with respect to the likelihood that your router might fail? I certainly don't know, but the fact that it has been running well for two years is a great sign! Why does there seem to be a bunch of cases being reported now? Who knows--given that many folks are now confined to working at home because of the COVID-19 outbreak, Internet access is perhaps more critical than at other moments in the past. Maybe for that reason, more people are confronting the issue head on? This is all complete speculation on my part!!!

 

[OllieNZ]

Good question--does this equally affect all R7800s from any time/date range of production? Or does it only affect a particular batch of R7800s? In the latter case, this could be explained, for example, by a batch of slightly less robust NVRAM chips that made it into a finite time/date range of production by Netgear.

Two things to keep in mind--this does not affect just the R7800... If you do a search on "netgear router loses settings" (https://www.google.com/search?q=netgear+router+loses+settings), just in the first page of results, you'll see lost settings mentioned in relation to a whole bunch of Nighthawk router models including the R7500v2, R6400, R7000, R7800, R7900P, R800P, and R9000 for good measure!!! Maybe they all use the same NVRAM chip (easy enough to figure out), and thus, maybe all can be pinned to a finite time/date range of production, which might indicate a batch of slightly less robust NVRAM chips?

Keep in mind that even though all those different models are mentioned with respect to lost settings, without reading any further, the specific symptoms and manifestation of the issue might be different from that which is discussed in this thread. For example, in some cases, it seems like folks were able to reload old configuration files, and quickly get back to business. But at least in my experience, and that of the others who have mentioned the issue in this thread, we were/are unable to reload old configuration files, which makes the problem super-annoying because it means having to painstakingly retype in one's whole configuration...

The other thing is that there seem to be a lot of hints that this issue is often related to power outages or failures--the NVRAM chips and supporting circuitry seem to be sensitive to this. For sure, this cannot be pointed to in all cases, but unless one is specifically next to the router when something like this happens, it might be hard to know. As I mentioned, with my new RMA swap R7800 that failed on Thursday, I was in the adjacent room when we had the power brownout, and I actually heard the UPS give one beep, and then, the R7800 went kaput. I suspect this also might have been what happened to my original R7800 a few weeks back, too, but I was simply not there to be able to confirm it.

So what does this mean with respect to the likelihood that your router might fail? I certainly don't know, but the fact that it has been running well for two years is a great sign! Why does there seem to be a bunch of cases being reported now? Who knows--given that many folks are now confined to working at home because of the COVID-19 outbreak, Internet access is perhaps more critical than at other moments in the past. Maybe for that reason, more people are confronting the issue head on? This is all complete speculation on my part!!!

Thanks for the insight, it has gone some way to convince me that I'm not sitting on a ticking time bomb. Luckily our power is pretty stable and I can't even remember the last time we had a power cut or even a flicker. I don't run a UPS or even a fancy surge protector so I'm probably courting disaster in general let alone with the router I think last time it got rebooted was when I installed the .74SF firmware.

 

[Alex Tzonkov]

Thanks for the insight, it has gone some way to convince me that I'm not sitting on a ticking time bomb. Luckily our power is pretty stable and I can't even remember the last time we had a power cut or even a flicker. I don't run a UPS or even a fancy surge protector so I'm probably courting disaster, in general, let alone with the router I think last time it got rebooted was when I installed the .74SF firmware.

I am in the same boat as you. My R7800 is over 2 years old and just recently began "freezing", the last time it froze, it came back with factory default settings. When I say freezing, the symptom for me was internet stopped working, and when I loaded up the Web GUI, all I see is the NG header and four radio buttons positioned in a rectangle ( I wish I took a screencap now). From reading this thread I have a couple of suspected things:

1. I recently (last ~2 months) added a USB3 drive and enabled and began using ReadyShare. (As @Voxel pointed out this causes unnecessary writes to NVRAM)
2. I also turned on traffic meter (about ~6months ago or s) and I never turned it off (Again I think this might have resulted in a lot of unnecessary writes to NVRAM)
3. Heat, with all the recent working from home, the temperature in the home office is higher with two adults being there every day for 2 weeks now, and keeping the ecobee temp higher as a result (This one is a bit of stretch but not zero probability)

My Router was always on Stock FW V1.0.2.52 I froze it there because of the issues with Intel wireless drives in subsequent Firmware versions, no idea if the driver was ever fixed. Here is what I am talking about, from the Release Notes on Stock FW V1.0.2.58

Known unfixed issues:

• Intel7260/3160 has a connection problem with the R7800 when the HT160 is enabled. This issue is due to the Intel driver. Intel is looking into this.

My current plan of action is to flash Voxel FW (just found it tonight) and stop using ReadyShare and look into enabling NFS access to the USB3 drive (by the way any pointers with that would be awesome)

Thanks you all, this thread was super informative.

 

[microchip]

@Alex Tzonkov

Get a cheap laptop cooler (preferrably with nice RGB LEDs ) and place it under the router to cool it down. It's what I do when temps get high

 

[Alex Tzonkov]

@Alex Tzonkov

Get a cheap laptop cooler (preferrably with nice RGB LEDs ) and place it under the router to cool it down. It's what I do when temps get high

Great idea thanks, will order one on Amazon Trying to find a nice small one for the R7800.

I am just hoping it is not too late and my NVRAM chip is not gotten "bad", will have to wait and see.

 

[LeKeiser]

Hello,
This is a very worrying problem, especially now that we, in France and in many other countries, are confined at home, and very very dependant on the network.
I own a R7800, very happy with it. I flash it with every new Voxel firmware when it is released. I guess I will stop for a while, in case of...
I always save my settings. I have a quite a lot of devices connected to the router. I'm always surprised to see how many, but that's how it goes. Some Google Homes, some Alexas, some iPads, phones, power plugs using the wifi, ...
I've been told that if I get a new R7800, I wouldn't be able to use those settings, I would have to configure every devices again. That sucks
My router sits in my bedroom. So far, it's not too warm. I have bought a fan for laptops last summer, it cools the router quite nicely. So far, I haven't turned it on. It's connected to the USB port of the router.

Well, crossing fingers now.
But you guys are scaring me :O

 

[tommytqt]

Hello,
This is a very worrying problem, especially now that we, in France and in many other countries, are confined at home, and very very dependant on the network.
I own a R7800, very happy with it. I flash it with every new Voxel firmware when it is released. I guess I will stop for a while, in case of...
I always save my settings. I have a quite a lot of devices connected to the router. I'm always surprised to see how many, but that's how it goes. Some Google Homes, some Alexas, some iPads, phones, power plugs using the wifi, ...
I've been told that if I get a new R7800, I wouldn't be able to use those settings, I would have to configure every devices again. That sucks
My router sits in my bedroom. So far, it's not too warm. I have bought a fan for laptops last summer, it cools the router quite nicely. So far, I haven't turned it on. It's connected to the USB port of the router.

Well, crossing fingers now.
But you guys are scaring me :O

Good idea with the fan. I looked online. They are so bulky. Is there any fan coolers only have 1 fan or slim or small profile?

 

[HELLO_wORLD]

Good idea with the fan. I looked online. They are so bulky. Is there any fan coolers only have 1 fan or slim or small profile?

I just ordered that https://www.amazon.fr/gp/product/B075WSD68T/
It is a glorified chassis fan with a USB plug. Overpriced but convenient and small (140mm). I will place it under the router, leaving some space under and above it.

I think @verbage is on to something with this temperature thing.
I should receive my replacement router today (UPS). I will be very careful with temps and writes to MTD flash memory.

 

[NetBytes]

I started with one similar to @HELLO_wORLD but using @kamoj's plug in I never really saw a drop in temperature in the router.
I got this https://www.amazon.com/gp/product/B01KZALHGK/?tag=snbforums-20 and my radio and CPU temps now consistently run 20 degrees F cooler than before without it.
Powered by one of the routers USB ports and perfect size to fit under the router.

 

[kamoj]

Comment:
Its possible to throttle the cpu.
That could lower temperature with at least 2 degrees.
The radios can already be set to lower effect, in the stock GUI.

 

[HELLO_wORLD]

I started with one similar to @HELLO_wORLD but using @kamoj's plug in I never really saw a drop in temperature in the router.
I got this https://www.amazon.com/gp/product/B01KZALHGK/?tag=snbforums-20 and my radio and CPU temps now consistently run 20 degrees F cooler than before without it.
Powered by one of the routers USB ports and perfect size to fit under the router.

Could not find this exact one here in France, but similar. Switched my order, thank you for the tip!

 

[verbage]

All, while I have mentioned heat as a particular potential factor, it is all complete speculation on my part. It does seem to offer a rational explanation of why--when a router is struck by this issue--that it seems unfixable immediately at the moment. But if you let it sit unplugged/unconnected for a few days (probably only a few hours would be necessary...), it seems to start remembering settings again.

But this is all based on limited observations from my own two R7800s, and potentially the additional one by @HELLO_wORLD in post #62 of this same thread (https://www.snbforums.com/threads/r...ettings-after-reboot.62773/page-4#post-565716). In my case, my two afflicted R7800s seemed to start working normally again after sitting unplugged/unconnected for a few days. If heat was the issue, this period of non-use would have allowed them to cool. In the case of @HELLO_wORLD's testing experience from post#62, it's not clear if his/her R7800 was left unplugged/unconnected for some time so as to cool down before it started remembering settings again, but it presume that is the case.

Another interesting observation from @HELLO_wORLD's post #56 is when his/her router started remembering settings again, it only lasted for a period of time before loss of settings occurred again:

Uninstalled Kamoj Addon, reboot... no factory reset!...Then I installed latest Kamoj 5.1b8, reboot -> factory reset!??

Again, heat provides a potential logical explanation--the router might have started functioning normally initially because it was cool, but after running for a while and warming up, it started exhibiting the symptoms when the Kamoj 5.1b8 addon installation was attempted. I am making several assumptions here!!! For sure, it will be better to hear from @HELLO_wORLD about this directly to understand if this interpretation might make sense.

So while heat seems to be a logical explanation, the reality is that it is only based on the few observations mentioned above!!! I.e. YMMV, and it is complete speculation at this point.

Whatever the case, this does seem to be hardware issue, and I suspect Netgear must know about it. But it is probably obvious why they are not necessarily saying anything--user J2793s on the Netgear forums hints about this in post #59 of the "R7800 Resetting to Default Config" thread there (https://community.netgear.com/t5/Ni...to-default-configurations/m-p/1888533#M153610). The suggestion is that they (Netgear) are probably fearful of a consumer class action suit...

 

[e38BimmerFN]

One thing user can do and should do for there routers there using, place a laptop cooler under there router. I've been doing this for almost 10 years now. There inexpensive and help keep the router cooler and helps extend life of the router. Heat over time will degrade electronics. Nature of the beast. I have my R7800 online now running Voxels FW and Kamojs add-on. Been working fine here for me.

 

[HELLO_wORLD]

I never left the router unplugged, so it never cooled down to room temp.
Usage now and before lockdown is similar on the router, so no recent sudden change. It worked perfectly fine until recently (reboots, etc...). So something changed (I suspect MTD chip to be failing).

the problem seems to appear more after several manipulations in a short period of time (installing fw, then the plugin), and it probably produces more heat than when it is left alone.

In my case, the router is not remembering, as each time it factory resets, settings are gone for good.

I think heat is not the main reason, but faulty MTD is.
However, heat seems to be a contributing factor to either damage the MTD memory faster or worsen the failures. Anyway, heat is not good for this (sensitive) router and cooling it down makes sense as well as reducing as much as possible writing operations to the MTD.

As for NG position, it seems clear that they are aware of the problem.
I asked for RMA with a link to NG forums about that issue, I was asked a few basic questions (to make sure of my issue), and it went up to engineers. A day later, they approved the RMA, without additional questions and without having me to proceed any manipulation on the router.

All, while I have mentioned heat as a particular potential factor, it is all complete speculation on my part. It does seem to offer a rational explanation of why--when a router is struck by this issue--that it seems unfixable immediately at the moment. But if you let it sit unplugged/unconnected for a few days (probably only a few hours would be necessary...), it seems to start remembering settings again.

But this is all based on limited observations from my own two R7800s, and potentially the additional one by @HELLO_wORLD in post #62 of this same thread (https://www.snbforums.com/threads/r...ettings-after-reboot.62773/page-4#post-565716). In my case, my two afflicted R7800s seemed to start working normally again after sitting unplugged/unconnected for a few days. If heat was the issue, this period of non-use would have allowed them to cool. In the case of @HELLO_wORLD's testing experience from post#62, it's not clear if his/her R7800 was left unplugged/unconnected for some time so as to cool down before it started remembering settings again, but it presume that is the case.

Another interesting observation from @HELLO_wORLD's post #56 is when his/her router started remembering settings again, it only lasted for a period of time before loss of settings occurred again:



Again, heat provides a potential logical explanation--the router might have started functioning normally initially because it was cool, but after running for a while and warming up, it started exhibiting the symptoms when the Kamoj 5.1b8 addon installation was attempted. I am making several assumptions here!!! For sure, it will be better to hear from @HELLO_wORLD about this directly to understand if this interpretation might make sense.

So while heat seems to be a logical explanation, the reality is that it is only based on the few observations mentioned above!!! I.e. YMMV, and it is complete speculation at this point.

Whatever the case, this does seem to be hardware issue, and I suspect Netgear must know about it. But it is probably obvious why they are not necessarily saying anything--user J2793s on the Netgear forums hints about this in post #59 of the "R7800 Resetting to Default Config" thread there (https://community.netgear.com/t5/Ni...to-default-configurations/m-p/1888533#M153610). The suggestion is that they (Netgear) are probably fearful of a consumer class action suit...

 

[verbage]

Another note--I will soon be able to contribute another data point about heat as a potential issue. On my new RMA swap R7800 that lasted only two days before exhibiting the loss of settings issue last Thursday, I mentioned that I had let it sit around unplugged for two days, and then on Saturday evening installed OpenWRT to see how it fared with that firmware. And that's when I recognized this second R7800 started remembering its settings again after being afflicted with the issue just two days prior. As I tested OpenWRT and got it all set up, I did several full resets and power cycles, and settings were preserved. But whatever the case, since we are supposing this is a hardware problem, it should equally affect any firmware*, right? I have not done any full reboots or power cycles on that OpenWRT-flashed R7800 since Sunday, and I am hesistant to try it right now for fear of disturbing family sanity with a lapse in connectivity during business/school hours. But I will try a full power cycle this weekend to see what happens.

* Though I make the assertion that a hardware issue should affect any firmware (official Netgear, Voxel, OpenWRT, or whatever) equally, this might not be 100% true. If the issue is also related to wear and degradation of the flash from heavy writing--like the Traffic Monitor constantly updating statistics--well, the assumption would hold only if all firmwares map/use the flash memory chip with the same footprint... But I do not think that is the case. Since Voxel's firmware is based on the official Netgear firmware, they must map out and split the flash memory chip with an identical footprint. By the way, that is how I had started with with my new RMA swap R7800. I immediately flashed it with Voxel's latest firmware, and, unfortunately, enabled Traffic Monitor on initial installation because I did not realize the issue of it constantly writing to flash). But I believe that OpenWRT maps out the flash chip differently...

So let's assume I already unfortunately wore out a portion of the flash memory in my new RMA swap R7800 because I enabled the Traffic Monitor. But that was under the Netgear/Voxel map of the flash chip. Since OpenWRT uses a different flash footprint, it may not use the segment of flash that has already been worn out.

Again, this is all complete speculation on my part, but if it were true, it might also allow a potential mitigation or solution to the problem for people who are already out of warranty. In particular, remap the flash chip to store settings in a different spot that is not worn out. The effect would be that you lose a certain amount of flash storage, but having a device that works normally again would certainly make this worthwhile. And then never enable Traffic Monitor again, or hard force it to log its statistics elsewhere, like on a USB drive.

 

[verbage]

As I was reading a thread about some flashing issues over on the OpenWRT forums (see https://forum.openwrt.org/t/r7800-flashing-openwrt-causes-bootloop-bad-block-in-kernel-area/49608), and saw a reference to the fact that different brands of NAND flash chips were used in different R7800s. Some reported a Micron-brand MT29F1G08ABBEAH4, and others reported a Macronix-brand MX30UF1G18AC. Nominally, these are identical components that can be substituted for another, but maybe one happens to be more robust than the other (instead of a just a lower quality batch of the same chip from the same company)? Maybe this explains why some have not seen this lost settings issue with their R7800 after years of use, but others have seen it relatively rapidly?

I just checked the dmesg from my current RMA swap R7800, and the NAND is reported as:

[ 1.038047] nand: device found, Manufacturer ID: 0xc2, Chip ID: 0xa1
[ 1.038079] nand: Macronix MX30UF1G18AC
[ 1.044615] nand: 128 MiB, SLC, erase size: 128 KiB, page size: 2048, OOB size: 64

Are folks with R7800s experiencing this lost settings issue using the same Macronix-brand NAND chip? And are those without this problem running with Micron-brand NAND flash? Perhaps not, but just something else to consider.

In the long run, if somebody's router is out of warranty, and no other software-based non-invasive solutions seem to work, I wonder if physically replacing the NAND chip would be a possibility? Just a quick scan of prices shows these chips are in the $3-$4 range. Certainly, the folks capable of this operation would be limited, but it might be preferable than dumping a whole R7800 just because the NAND flash chip is defective...