svkatielee

Thanks zador. When I asked the question the kernel was in the repo, but none of the images were updated on the download page so I jumped to the wrong conclusion. Just being impatient. haha

Do I understand this correct? The 5.20 reflects the patch level of the kernel, not necessarily anything about the "Armbian distribution". Such as when download images will have NM instead of wicd. Or maybe you meant: the kernel that will be in "Armbian distribution release 5.20" is now available. I just did the upgrade: The following packages will be upgraded: linux-firmware-image-sun8i linux-headers-sun8i linux-image-sun8i linux-jessie-root-orangepiplus2e linux-u-boot-orangepiplus2e-default and the apt-get dist-upgrade says nothing to do. This is: cat /etc/armbian-release # PLEASE DO NOT EDIT THIS FILE BOARD=orangepiplus2e BOARD_NAME="Orange Pi+ 2E" VERSION=5.20 LINUXFAMILY=sun8i BRANCH=default ARCH=arm Boot and login seem faster, but was already nice on the 2E installed to th eeMMC. Thank you. Larry

I set static IPs for several Opi's with nn-applet seems to work ok. However, I had to do it al as root. My user can not edit connections still. Armbian Jessie 5.18 from above link. No desktop, only nodm. if I su - root then nmtui I can edit oK. any other, the edit button is grayed out or the OK button is.

Thank you. It seems to work fine. I still have not had time to figure out hostapd with NM yet. My other life gets in the way. Where is the additional discussion where you have decided to go ahead with switching these out instead of wicd? I have been having trouble finding the discussions since the big documentation change.

See this post for probable problem: http://forum.armbian.com/index.php/topic/1775-wicd-gets-erroneous-bad-password/?p=14708 Since the problem seems to be a wicd to wpa_supplicant timeout problem, any change in information (password, driver, SSID,etc.) can change the amount of time it takes wpa_supplicant to reply. So lots of these differing solutions work some of the time.

I've just spent 2 complete days trying to reproduce the above steps for switching from wicd to NM and failing. (with 5.16 Opi 2E eMMC) If I run vncserver as root, nm-applet works just fine: connects to wifi, edit connections. But not as a user. Finally, the difference, I was trying to do it with tightvncserver (VNC) instead of a direct monitor. In frustration this morning I plugged the monitor in and there was the dialog pop-up to unlock the keyring! So these are not the complete instructions for using VNC exclusively.

Where is the final destination for the HOWTOs? Since this re-organization of the github vs forum, I can't find the discussion streams about the claimed tasks.

@Igor I wonder if you had as much trouble getting NM installed as I did? I tried for a combination that did not need gnome-keyring and network-manager-gnome because of all the extra support packages. And I did not want to enter a password at every boot. To avoid the password I had to use the null passphrase in the default keyring which I thought might not be acceptable. I also tried to find a different alternative. I didn't find any that seemed to work headless and in a desktop. Most choices were various subsets of wicd or NM (wpa_supplicant, netctl, etc.). I even tried to compile wicd.1.7.4 but it needs things in newer kernels I think.

It has taken a while but finally I now have Network Manager running instead of wicd. Here is what I did on my Opi PC 2E installed and booting from nand: Armbian Jessie desktop 5.14 apt-get upgraded to 5.16 and updated Aug 10, 165. Connected to cheap CDR-King Wireless-N AP Router (Chinese) apt-get install network-manager network-manager-gnome xfce4-notifyd xfce4-indicator-plugin gnome-keyring gcr libgck-1-0 libgcr-3-common libgcr-base-3-1 libgcr-ui-3-1 p11-kit p11-kit-modules apt-get remove --purge wicd python-wicd wicd-daemon wicd-gtk gpasswd -a USERNAME netdev Comment out eth0 from /etc/network/interfaces, (also wlan0 if you have it) Change managed= to true in /etc/NetworkManager/NetworkManager.conf Create new file with this content: /etc/polkit-1/localauthority/50-local.d/org.freedesktop.NetworkManager.pkla [nm-applet] Identity=unix-group:netdev Action=org.freedesktop.NetworkManager.* ResultAny=yes ResultInactive=no ResultActive=yes with eht0 connected and on-board WiFi, reboot came up with eth0 and WiFi available click nm-applet in menu bar select ifupdown (eth0) Ethernet comes up selected my wifi ssid Prompt for password entered password, prompted for default keyring password, left blank and click Continue and Continue again for clear text storage of password rebooted, still have both connected, but icon in nm-applet only shows WiFi but right-click->Connection Info shows both. Now for some testing ....

I am willing to try some. I have only Orange Pis: One, PC, PC +, and PC 2E. I have a few cheap wifi dongles and a couple really cheap Chinese routers. What I don't have is a good understanding of hostapd/network manager and all the wpa_supplicant pieces. And I don't' have much network access. I can't run your build tools because the network is too slow and unreliable. So with a little guidance I'll give it a shot.

I found an old bug report on the same version of wicd that is delivered with the legacy builds. Try to authenticate WPA for wifi, get bad password error. Repeated attempts to authenticate as many as 10 times before it is successful. It will finally succeed, therefore; the password is not bad. https://bugs.launchpad.net/wicd/+bug/1084380 A short summary includes, there seems to be a problem in the wpa_supplicant that causes it to occasionally take too long to return from an authentication request. wicd rev 1.7.2 is set to wait 35 seconds then times out and reports BAD PASSWORD, when it should have reported a timeout. The suggested fix was to allow user config for the timeout, but the test of 250 system startups showed an averaged between 35 and 55 seconds. Their request was to set the default timeout for wpa_supplicant to 120 seconds. The fix was promised for 1.7.3. Jessie/Trusty have wicd 1.7.2.4-4.1 Xenial has 1.7.4.xx I believe this is why so many complaints are posted about wifi on the various armbian SBCs. And why there are so many different "solutions" that only work in special cases. But this is a work-around for wpa_supplicant taking too long in some instances. Can we get a newer release of wicd in the build? I tried just downloading and installing wicd 1.7.4 but without anything else, it no longer connects at all ....

I am a sailor and use OpenCPN for the navigation chart plotter. The charts load and zoom much faster with hardware openGL/ or ES. It has a way to display rendering technique and frames/sec and is barely useful without acceleration. It is very nice with it, even on the Orange Pi One. But I will use the Opi PC Plus for my final installation. I compiled opencpn(4.2 and 4.4) on the default Jessie desktop of both 5.05 and 5.10. I will be recompiling on 5.17 soon. So it is not only games and display managers that use openGL and hardware acceleration. Thank you all for armbian and the rest.

I have just discovered the u-boot help file. On console during power-up, when it says Hit any key to stop autoboot: I hit a key, got a prompt and typed help. Are all the capabilities listed real? pxe, tftpboot, nfs, usbboot, dhcp, bootefi Are many implemented? I did just discover the nfs how-to. -- Larry

Sorry I was not complete and accurate. Debian Jessie desktop 5.14 apt-get upgraded to 5.16 as of 36 hours ago. The gui in the upper right of the default X interface autologed in as the user defined during install The Wicd Network Manager GUI, properties of the wireless AP, selected "WPA 1/2 (Passphrase) preshared key" The first key was "bbbbbbbb" which could be interpreted as HEX or as ascii. The second key was "katielee" unmistakably ascii. The bbbbbbbb was changed to bbbbbbbbz and then it worked.(did not get "Bad Password" and succeeded in connecting to network.) As for code I meant the probable C code that the gui is using to evaluate the key. The log of my install is here: https://www.facebook.com/groups/1483270805309190/permalink/1573610346275235/

Same problem, but this did not help. Tried two Opi PC pluses on two different networks, two different sd cards. It seems to be related to the password. Both networks are set to WEP (Passphrase), but have different passwords, each only 8 characters. One password is too simple, all characters could be either hex or text. the other network's password is also 8 characters but random alpha. We altered the simple password network password by adding a 'z' to the end of its password and now it works fine. It seems as if the wireless code is having difficulty between HEX ans Passphrase.

Nice job on the Documentation. Is this where I could give feedback? 1. It would be really helpful if you included a mechanism to provide "info valid as of date" along with each subtitle or at least each page. The entire internet has forgotten the importance of dating information. For instance: Build Process -> Desktop installation : hardware acceleration only... If it is still true, will someone remember to update this when it changes? 2. It feels like the title is missing for some pages. Build Preparation , Build Options, User configurations 3. Apparent formatting mistake in table at: Hidden options for advanced users (default values are marked bold) USE_CCACHE - entire line is bold FIXED_IMAGE_SIZE - all in left cell

Be careful here, hostname -i (lowercase) reports whatever is in /etc/hosts no mater how erroneous it might be. I had several optional names listed in hosts for cellular (2G and 3G), wifi and wired. (Ubuntu 15.04) With the capitol -I it does report only the active one(s). My VirtualBox VMs report the pretend link to tthe host as well as the private one direct out to the net. (Ubuntu 14.04, MintMate 17.1) However my Orange Pi PC (5.10 upgraded to 5.13) reports only the active address for each conmmand option. Even with multiple entries in hosts. None of the systems reported 127.0.0.1 (loopback) with either option.

I am trying to build an image for both the Orange Pi PC and One I am on a slow cellular network and pay by the gigabyte for access. Is there a way to cache the packages for: Installing base system [ Stage 1/2 ] , and all the rest? Each time I try a compile.sh it take hours to re-download the whole set of packages, because the there is almost always at least a few packages that could not be fetched due to network errors, then they are unpacked and a corrupt tar image is saved and the packages are removed. If I save the create_rootfs_cache file with the missing packages, it does not try to reload them, it just errors out cause they are missing. I am building in a Ubuntu 14.04 VM on a Ubuntu 15.04 host. I have trird various things in compile.sh, currently it is: BOARD=orangepione PROGRESS_DISPLAY=plain PROGRESS_LOG_TO_FILE=yes RELEASE=jessie BRANCH=default NO_APT_CACHER=yes And this last attempt my userpatches/lib.config contains only: DEBIAN_MIRROR="ftp.tw.debian.org/debian/" because I am in the Philippines. Larry Edit: And this pass failed to after 2 hours; I: Retrieving zlib1g 1:1.2.8.dfsg-2+b1 I: Validating zlib1g 1:1.2.8.dfsg-2+b1 E: Couldn't download packages: libpipeline1 cp: cannot create regular file ‘/home/larryl/armbian/output/cache/sdcard/usr/bin/’: No such file or directory [ o.k. ] Installing base system [ Stage 2/2 ] chroot: failed to run command ‘/bin/bash’: No such file or directory [ error ] ERROR in function create_rootfs_cache [ debootstrap-ng.sh:163 ] [ error ] Debootstrap base system second stage failed [ o.k. ] Process terminated

If you notice above, I did try SSH, multiple times. Thanks, Can you expand on the additional tuning options available (with SSH)? I'm not too good with that. I would have particularly liked an option to 'not erase everything' on failure. Once it managed 114 out of 148MB and several times as much as 20Mb. If it hadn't erased each try I would have been done a couple days earlier.

If you notice above, I did try SSH, multiple times. Thanks, Can you expand on the additional tuning options available (with SSH)? I'm not too good with that. I would have particularly liked an option to 'not erase everything' on failure. Once it managed 114 out of 148MB and several times as much as 20Mb. If it hadn't erased each try I would have been done a couple days earlier.

I finally succeeded!. I did the git clone to a friend's system in Seattle. Then did an scp back to the Philippines. It worked. The compile has no errors. Soon I'll try to boot it and see if the DisplayLink USB monitors work. Github still does not like the local network, but it can succeed in a pull, if not much has changed. Thanks for letting me rant.

I have been trying to compile a legacy kernel for Orange Pi PC for 48 hours. (I want to add the DisplayLink driver for USB monitor like the Lilliput 10") I have a very poor network connection , wifi to my boat in the marina work yard, Philippines. Then I tried each of these via the cell phone network (my phone as a hotspot) ./compile.sh fails so I tried cloning the repo by cmdline, and at github and ssh vs http: git clone -n https://github.com/igorpecovnik/linux.gitsun8i -b sun8i --depth 1 git clone -n https://github.com/svkatielee/linux.gitsun8i -b sun8i --depth 1 git clone -n git@github.com:svkatielee/linux.git sun8i -b sun8i --depth 1 Each time it fails, git deletes everything it did download, so I can't get there incrementally. The most common error: error: RPC failed; curl 56 SSL read: error:00000000:lib(0):func(0):reason(0), errno 104 fatal: The remote end hung up unexpectedly fatal: early EOF fatal: index-pack failed Then I tried all the combinations of Download Zip I get any where from 2 to 100 Mb then it hangs and fails, but usually less than 10Mb. Then I went to a friends boat, 6 Mb before Download Zip froze then errored out in Firefox (where I don't see the message) Is there any other place to download a zip file of linux sun8i that is more resilient to dropped packets and congested network? Is there a way to break up the repo into smaller pieces that I might eventually get downloaded? Maybe a torrent? Is there a problem with github.com?

I have compiled OpenCPN Marine Chart plotter and navigation software for the Orange Pi PC with the Mali graphics acceleration. It seems to be working fine. I am trying to document the process and possibly make a prebuilt image available at opencpn.org. http://opencpn.org/ocpn/Install_Orange_pi_Allwinner I am not happy with the name they gave me for the page. And the details at: http://kb7kmo.blogspot.com/2016/04/opencpn-on-orange-pi-allwinner-h3-arm.html My question is how should I reference armbian? And how should I reference the hardware? Since it appears that a number of these boards are very similar and would probably require the same steps on the same version of armbian? I only have the OPi PC and OPi One. Thanks, Larry

I found Igor's Mar 12 commit, https://github.com/igorpecovnik/lib/commit/bf66a46085bdd8d6abc6149c710e0c0532dc882b Automount/eject of USB, fix reboot/shutdown on desktop (Jessie) So I followed it: sudo apt-get -y install gvfs policykit-1 policykit-1-gnome eject edited /boot/boot.cmd to change "cgroup-enable" to "cgroup_enable" sudo mkdir -p /etc/polkit-1/localauthority/50-local.d/ added plugdev.pkla and power.pkla to it plugdev.pkla [Allow Automount] Identity=unix-group:plugdev Action=org.freedesktop.udisks2.filesystem-mount;org.freedesktop.udisks2.filesystem-mount-system;org.freedeskt op.udisks2.filesystem-mount-other-seat;org.freedesktop.udisks2.filesystem-unmount-others;org.freedesktop.udis ks2.eject-media;org.freedesktop.udisks2.eject-media-system;org.freedesktop.udisks2.power-off-drive* ResultAny=yes ResultInactive=yes ResultActive=yes power.pkla [Allow Desktopstuff] Identity=unix-group:sudo Action=org.freedesktop.login1.reboot;org.freedesktop.login1.reboot-multiple-session;org.freedesktop.login1.po wer-off;org.freedesktop.login1.power-off-multiple-sessions;org.freedesktop.login1.suspend;org.freedesktop.log in1.suspend-multiple-sessions;org.freedesktop.login1.hibernate;org.freedesktop.login1.hibernate-multiple-sess ions ResultAny=yes ResultInactive=yes ResultActive=yes I still get an error window saying I do not have permission to eject, however the USB stick is indeed unmounted. So fixed. Thank you for your comments. Larry

I think not. larryl@orangepione:~$ groups larryl dialout sudo audio video plugdev netdev Does your USB stick automatically mount (or thru thunar) when pluged in? If this is my (and Seasalt's) personal problem, I'll live with it. I assumed it was widespread and that I would be helping to figure it out. I have orangepione, 5.05 image on 32Gb Samsung EVO, sudo apt-get update, upgraade added Cyphers to sshd, apt-get install tightvncserver xtightvncviewer, No other packages loaded. Before on the Opc PC, I installed gvfs policykit-1 and later removed policykit-1 thinking that was the problem, but not installed this time.