Joined: 31 Dec 2014 Posts: 9 Location: Huntington Beach, CA
Posted: Fri Feb 20, 2015 23:46 Post subject: Re: Issues with /jfffs setup
jheide44 wrote:
jtaitt wrote:
...
I believe I have found a bug in setting up /jffs. ...
when I setup /jffs, the clean only lasts for a few seconds and when looking at the contents of the jffs folder I find it with the files and folders from the last setup.
...is there a CLI or Shell command which can be used to clear this area of memory outside of the GUI
what are you mounting to that /jjfs
a USB HDD/USB Flash drive/or just on the internal memory?
and what files do you see? (do files files you copied/created persist through the cleaning?)
i.e. given your configuration say if you had say minidlna enabled you should expect to see a minidlna directory with files.db file in side it (automatically put there by your device)
there are unix like CLI commands to erase files & directories, but I just wanted to ask a few questions before sharing anything more
I'm setting jffs up in internal memory only. Initially the folder was empty. I added 3 other folders, bin, etc, web and a dnsmasq.custom file which contains a lot of static leases I want to kept and not have to rekey after each flash as I would storing them in nvram by keying them in the GUI. Later I changed my settings to get into the GUI from http to https. The system generated the SSL files and placed them in the jffs folder. Somewhere along the line the system also created a tmp folder which I looked into and it contained most if not all of which is in the /tmp folder off of "/" root, which I assume is used by the system itself.
After each Flash of a new build and again setting up the jffs folder all of the same files and folders are still there even when I have tried to clean the folder as it is setup in the Gui.
After loading 26285M I went and found the same files, I manually deleted everything in the jffs folder and added back my dnsmasq.custom file. I again, changed http to https and the SSL files did not come back. The system created a tmp folder in the jffs folder again but now it is empty.
Weird stuff! I have backups of all of the new builds I have tried since the new drivers were added. I want to go back and look at them and verify what I am seeing as well as clear out all memory, nvram and the memory where the /jffs folder is stored and then flash the newest build and completely rebuild it from scratch and see what happens to verify my findings. They may have been in earlier builds and have now been fixed.
Thus the reason I didn't want to go into a lot of detail before I was sure there was really a problem. Perhaps the initial clean of the /jffs setup is a one time occurrence and in later builds it never gets cleared and when you try and clean it again it only takes a few seconds because it sees the folder as already there and doesn't erase it and start over, It may be by design. But so far what I have read this doesn't sound like what it should be doing, and that is wiping it clean, reformatting the space and what ever the clean process did on the first time I set it up.
Hope that makes some sense? Thanks for your interest... _________________ Netgear r7000 26450M
I been getting hounded by the members on my network coz the network hasn't been as stable as it was with old drivers.
Each build squashes old bugs and adds new bugs at least for me this has been the case for the past few builds, but then again this is part of the game.
I went back to 26160 and finally got everything working and no speed issues from what others have said hopefully.
Which firmware did Kong say we had to reset after flashing? I think it was 26160 but just want to avoid resetting if I can when I upgrade.
PS I think erase nvram over CLI for sure helped in my case.
Has anyone else encountered or knows what is causing these errors? Occurs on both latest BS and KONG builds on an R7000. KONG? BS? Anyone else?
MDT stands for memory technology device http://wiki.openwrt.org/doc/techref/mtd MTD devices represent flash devices and they consist of eraseblocks of rather large size, typically about 128KiB. Block devices consist of small blocks, typically 512 bytes. MTD devices support 3 main operations - read from some offset within an eraseblock, write to some offset within an eraseblock, and erase a whole eraseblock. Block devices support 2 main operations - read a whole block and write a whole block.
Observation/Questions:
If others are not reporting the same errors during boot, it could be something specific to your flash or memory chip. specifically the locations/addresses reporting the errors. (think about what his happening during boot sequence between volatile & non volatile memory... at the hardware level)
other than the errors noted in the image above are you experiencing any tangible "issues"?
has anyone else noticed these errors?
if you aren't having any issues, and nobody else is reporting similar errors.... I'd ignore it. < -- but i'm just some crazy guy from the internet!
disclaimer:
I reserve the right to be wrong & misinterpret the world around me. My experience is not comprehensive & based on other types of projects.
as far as file systems... UBIFS is the successor of JFFS2, I know a *little* about UBIFS... less about JFFS2.
All came from playing around with Android/NANDroid flash backups... mostly.
Of course this is a wild ass guess, feel free to correct me if i'm "out in right field picking dandy lions".
jheide44,
Thanks for the info!
There are no "perceivable/tangible" issues per say, however anyone who isn't running a log program for saving/displaying logs probably isn't seeing this error, and no doubt there are probably others who have it too, but just don't see it and also don't seem to have problems.
I just get nervous anytime I see any sort of CRC error when it comes to storage.
I highly doubt it's flash going bad or anything like that... I've only had this router about 5 months, and flashed it maybe 16 times over the course of that time. I'm sure the devs probably flash a single device that many times a month, if not more...
I'll just keep an eye on it, and if it suddenly stops reporting those errors in a future build, then I'll report back and check the SVN to see if a recent commit did anything flash/storage related. _________________ Netgear R7000 w/r31780M <KONG> build
Netgear R6700 (Un-opened with stock. My backup/emergency router if the R7000 takes a dump...)
2x Buffalo WHR-HP-GN 28493 (Used for 2.4 Ghz bridge when needed.)
Asus WL-500g Premium (1x v1 & 1x v2) (Still have, but retired for now.)
1x Linksys WRT54G v8 >>DD-WRT v24SP1 (The other routers needed something to point at and make fun of.)
@jebise101
FWIW I've been nothing but rock solid on BS 26138 and Kong PTB 26135, spent the bulk of the time running the BS build though. Haven't kept up with the Kong PTB builds though (been off site & away from my toys most of this month)
the wife and kids noticed the bug fixes cause we were able to turn on miniDLNA again without any "mysterious" reboots
@jtaitt
from initial question just wanted to make sure you knew up vs down before talking about RM from the CLI...
I see you are operating at a much higher level now (sorry) just didn't want to give a new CLI user a weapon of mass destruction.
@SirSilentBob
SirSilentBob wrote:
...<long boring conversation removed but it is up ^^there^^ if you are interested>...
There are no "perceivable/tangible" issues per say, however anyone who isn't running a log program for saving/displaying logs probably isn't seeing this error, and no doubt there are probably others who have it too, but just don't see it and also don't seem to have problems.
...
fwiw, i'm on a R6300v2, but i think we have the same flash chip... i'm running 26138 also (obviously not the same pre-compiled .bin file from main svn repo given our device level sub directories)
SSH'ed over to her, enabled jffs2 and Syslogd then rebooted
when she woke up again and let me back in asked grep to check the log for 'jffs' and 'error' results detailed below, but I didn't see any errors in the log.
you can see from mount what else I've got going on a couple of places are pointed at /jffs (but that shouldn't matter, right?)
was not local so couldn't pull the usb drives for this quick & dirty test
**If my logic was messed up and this test is not valid please point me in the proper direction (i.e. didn't run a special program for logging/review)** _________________ reenignE esreveR
Resistance is NOT futile... It's Voltage divided by Current!
Assumptions:
1. Everyone on the forum has read the relevant forum section announcements.
2. For Broadcom section we have ALL at least tried to understand the "Peacock" thread,HERE
Joined: 31 Dec 2014 Posts: 9 Location: Huntington Beach, CA
Posted: Sat Feb 21, 2015 4:48 Post subject:
jheide44 wrote:
@jtaitt[/color]
from initial question just wanted to make sure you knew up vs down before talking about RM from the CLI...
I see you are operating at a much higher level now (sorry) just didn't want to give a new CLI user a weapon of mass destruction.
No reason to be sorry jheide44, in hind sight my question wasn't really clear. My thoughts were since the clean option for jffs was not working as I expected, I was hoping there was a command like erase nvram that I could use rather than using rm -r as root to clear it all out. Especially if my previous command was cd /, then I would be taking a hammer to my router. LOL
And also I'm a little new to DD-WRT and the way I have seen things set up under the hood so to speak are a little different than a linix file or web server. So I don't want to cry out "BUG" until I am really convinced I'm seeing something which I think is a possible issue. And I've been a programmer for 30+ years and like to know what is going on under the surface. Like a little kid I like to take things apart to see how they work.
All to say I want to go back and redo the last 3 builds I did and then go through them again to make sure I saw what I think I did...
I do appreciate all of you guys who field all of the questions in this forum. There is so much information out there which is out of date, it makes some rather simple tasks seem really difficult at times and having people around who are up with the new changes really helps... _________________ Netgear r7000 26450M
Thanks for the follow up. I also grepped and viewed the messages file, and curiously enough, I'm not seeing those errors in the logs either. It could be that since I'm also logging dropped and rejected connections, that I've made the log data "roll over" and isn't showing the info from when it boot last (I'm certainly not seeing anything from when it was boot).
Output of mount is as follows:
mount
rootfs on / type rootfs (rw)
/dev/root on / type squashfs (ro,relatime)
proc on /proc type proc (rw,relatime)
sysfs on /sys type sysfs (rw,relatime)
debugfs on /sys/kernel/debug type debugfs (rw,relatime)
ramfs on /tmp type ramfs (rw,relatime)
none on /dev type tmpfs (rw,relatime,size=512K)
devpts on /dev/pts type devpts (rw,relatime,mode=600)
devpts on /proc/bus/usb type usbfs (rw,relatime)
/dev/sda1 on /tmp/mnt/sda1 type vfat (rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=utf8,shortname=mixed,errors=remount-ro)
/dev/mtdblock/5 on /jffs type jffs2 (rw,relatime)
It looks pretty similiar to yours, it looks like the only difference is sda1 on my unit is the flash drive that's currently plugged into the front of the unit.
Thanks for your wisdom, and I'll consider this a "non-issue" unless I actually encounter real issues with reads/writes from /jffs
Posted: Sun Feb 22, 2015 15:06 Post subject: Re: newd2 drivers bug
falcon111 wrote:
newd2 broadcom drivers have one more problem.
periodically wireless (5ghz) clients are loosing internet connection. signal strength is acceptable, but no traffic is being transferred via wifi. can't even ping my router - no ping.
If i look at the router from the wired connection, there is quickly increasing number of errorneous transmitted packets.
Other wifi clients still working ok.
On the dropped off client i have to disconnect from the router and then connect back again, and everything works ok. until next time one of the wireless clients will stop working.
This started after i've flashed from 24345oldd to latest newd2 builds.
I'm seeing the same thing. I've done the complete reset when I upgraded to build 26190 as well. I'm using the same 5ghz settings I used on the old drivers with months uptime. The router is fine over wired and 2.4ghz, it's only 5ghz that stops broadcasting completely. The SSID is nowhere to be found. Simply restarting the wifi brings it back up
Try resetting via the command line instead of the gui, the reconfigure.
It has corrected some folks odd behavior. _________________ Router currently owned:
Netgear R7800 - Router
Netgear R7000 - AP mode
Posted: Sun Feb 22, 2015 18:47 Post subject: moving from tomato?
Any special considerations when moving from tomato? This is what I will do:
Reset to default with tomato. Install dd-wrt from the web-gui again with reset to default. After dd-wrt boots, telnet with default password and do nvram erase and configure stuff.
Posted: Sun Feb 22, 2015 19:57 Post subject: Re: moving from tomato?
oh no, the current main SVN reposity pre-compiled REV# is higher than the currently posted Kong Personal Test Build (2/22/15 vs. 2/16/15)... what ever shall we do?!?!
...just kidding...
**crunch** <--- the sound of gears changing
csbvpx wrote:
Any special considerations when moving from tomato? This is what I will do:
Reset to default with tomato. Install dd-wrt from the web-gui again with reset to default. After dd-wrt boots, telnet with default password and do nvram erase and configure stuff.
--
Thanks.
Edit: in a few posts Malachi will point out what the best practice is. to avoid confusion I've edited this becase he is correct (& i was wrong)
From tomato return to stock (using their process/documentation). then from stock firmware go through the initial DD-WRT flash process.
It will put all the variables back where they are supposed to be, which is what DD-WRT expects.
sorry for my original/bad advice below...
All credit to Malachi
BAD PRACTICE:
FWIW your logic seems sound. I've flashed between Shibby Tomato & DD-WRT back in the E2500 days without issues.
If you have trouble logging in just remember the previous admin usernames and passwords, you should be fine (assuming nothing new in these ARM builds)...
Let us know how it goes. _________________ reenignE esreveR
Resistance is NOT futile... It's Voltage divided by Current!
Assumptions:
1. Everyone on the forum has read the relevant forum section announcements.
2. For Broadcom section we have ALL at least tried to understand the "Peacock" thread,HERE
Last edited by jheide44 on Sun Feb 22, 2015 23:02; edited 1 time in total
Joined: 28 Jan 2015 Posts: 41 Location: North Germany
Posted: Sun Feb 22, 2015 20:27 Post subject:
The new BS build is running fine on my RT-AC68U. I did a full reset before and after. Anyway the language support is still broken, nothing but english works. Anybody else having this issue?
The new BS build is running fine on my RT-AC68U. I did a full reset before and after. Anyway the language support is still broken, nothing but english works. Anybody else having this issue?
EDIT: style selection without function as well
post your findings over at the "New build is out 26332 (BS) K2.6 and K3.X" thread too...
I'm at IKEA today. the wife forced me to go shopping, just short of being dragged away at gun point (we are both CCW permit holders, lol). the time I was hoping to tinker with my toys this weekend has been seriously encroached upon
I'll report in with flash results and initial testing on the new build once I get a few moments of peace. ...but I have the sinking feeling I will be assembling furniture for the foreseeable future...
you know what they say, "happy wife, happy life" _________________ reenignE esreveR
Resistance is NOT futile... It's Voltage divided by Current!
Assumptions:
1. Everyone on the forum has read the relevant forum section announcements.
2. For Broadcom section we have ALL at least tried to understand the "Peacock" thread,HERE
