Update: At present I am trying to bring up the most recent FDT version on that system's SSD since its firmware has already been flashed to the 2017.07 version. That would hopefully solve the boot problems without doctoring up the non-FDT system drive. If that fails I'll go back to the non-FDT system and fiddle with it, but as bodhi has noted, non-FDT systems are becoming less thaby BuckNaked - Debian
Thanks, bodhi. I owe you another beer. Do you suppose this is a topic that should go into the wiki, or have non-FDT kernels been abandoned by everybody else?by BuckNaked - Debian
First off let me apologize for dredging this stinkin' issue up again. But, it seems that every time I touch one of my E02s, something goes wrong and over the last five years or so, it's always been the same thing - it hangs loading the kernel. And if you say it's my fault because I'm using old non-FDT systems, that's probably true. There's never enough time to updby BuckNaked - Debian
Well, I had to plug into it for several days while I was trying to figure out what was going on, so I figured a temporary solution wasn't a good option. After clearing the TX and RX vias, I soldered Arduino Dupont-style jumpers (pin ends) into the vias and grounded to the solder pad on the side of the SD socket. This gives me decent quality pins to plug into the USB serial adapter'sby BuckNaked - Debian
QuoteBodhi My advice is also consider a Sandisk USB thumb drive. Makes good sense. I was using a PNY and assumed that since it booted everywhere else, it would also boot in the Pogo V4. So much for assumptions.by BuckNaked - Debian
After several hours of exploring possibilities, I have a solution. I don't know if it will be of value to anyone else but here it is. This Pogoplug V4 will not boot from a USB drive. It will only boot from an SD chip, and it is particular about which of those it will accept. Despite the USB drive being readable and bootable on every other system, the Pogo V4 will not read from it. Aby BuckNaked - Debian
Bodhi, always good to hear from you. You should certainly be paid more for this job. Here's the information. I do not know why e2label fails claiming disk corruption. fsck reports that the disk is clean. The 4 GB USB pen drive being used for the boot is the same one I use for doing rescue operations on the E02. I cleaned off the Debian system (tried both systems first, one FDT andby BuckNaked - Debian
I'll start by saying that whoever at Cloud Engines eliminated the "Enable SSH" option should be staked out in the warm, friendly sun to concentrate their sugars for the benefit of our hard-working friends, the ants. Connecting to the serial header on the V4 is a colossal annoyance. Then I went through three USB-to-logic-level serial adapters before finding one that the Pogoplug wby BuckNaked - Debian
I've written an LED daemon for Pogoplugs running Debian. The daemon flashes the front panel green LED in either International Morse code or ASCII to display a text message. Display speed is settable from 5 words per minute to 30 WPM. It was written primarily for entertainment but may serve a useful purpose on systems that are not monitored frequently. If the system locks up and the daeby BuckNaked - Displays
Expanding on renojim's solution above ... the problem occurs due to linux file buffering / caching. It would be possible to get around it by calling fflush after every fputs, but the resulting code would be ugly and I believe that using write is a better solution anyway.by BuckNaked - Debian
That works great. I'll adopt the code into the original program. (Is it a disgraceful abuse of 70 years of hard work, technology and computer science to send Morse code through an LED on a Pogoplug?)by BuckNaked - Debian
I'm chasing what I think should be an obvious issue, but I can't find it. The host is a Pogoplug E02 running Debian jessie. Logged into the system as root, the following two commands turn the green LED on and off as expected. root@www:# echo "default-on" > "/sys/class/leds/status:green:health/trigger" root@www:# echo "none" > "/sys/clby BuckNaked - Debian
Thanks for the info on the supply. Yeah, I think one of the electrolytics dead-shorted. Electrolytics aren't a long-lifetime component, but they're way cheaper than tantalum ... and if it saves 15 cents, in an item with an expected five year life, cheap wins.by BuckNaked - Off-Topic
The power supply in my E02 failed catastrophically. Transformer is a melted blob. Bad cap, probably. After checking the mainboard in another unit, I confirmed that the mainboard is still operational. If anybody has a scrap Pogo with a working power supply -- I believe the E02, P01, P02, P21 and P22 all use the same supply -- and would be willing to part with it, please let me know. I knowby BuckNaked - Off-Topic
Thanks for the info. It's helpful to know that something has been done successfully before, so that if problems do arise there is no "maybe this doesn't work" hovering in the shadows.by BuckNaked - Off-Topic
Does anyone have experience with booting a Pogoplug E02 from an SSD via a USB to SSD adapter? (typical adapter on fleabay) Background: I've always been suspicious about running a production system (in this case a web server) from a flash drive. When flash drives were SLC I felt it was tolerable, but that suspicious feeling has become more so now that nearly all flash drives are MLC wby BuckNaked - Off-Topic
Bodhi, thanks for your assistance through this issue. It appears that there is no interest in the problem at the debian forums. Submitting it as a module-assistant bug might produce results later this year ... which is too speculative and too far out to be a solution. DKMS does not appear to be a solution. I think I will need to build a custom kernel. (a) Is there a tarball for the 3.17by BuckNaked - Debian
Have posted the problem to the debian forums at http://forums.debian.net/viewtopic.php?f=5&t=120128&sid=d05c351d5e802d3d46459ba14e407b59. I will post back here if anything develops. I believe you are correct about the problem being in module-assistant. The maintainer's comment "I've long ago switched from module-assistant to DKMS, so I don't really care what happeby BuckNaked - Debian
So far: module-assistant is a Perl script the problem occurs in the "prep" subroutine turning on the -d debug option didn't produce illumination of the issue and here is what I have seen so far after inserting some debugging trace prints near the problem area: 561 sub prep { 562 print "chdir $usrc\n" if $opt_debug; 563 %envbackup=%ENV; 564 $ENV{ROOT_CMby BuckNaked - Debian
Bodhi, thanks for the suggestion. The situation improved somewhat after installing kernel-package. I had hacked up the system so much by that point that I started again from scratch with a clean install of the rootfs. I reduced the commands to: ** Install the kernel headers from the install kit ** cd /usr/src wget --no-check-certificate https://bitly.com/1zrgVzP mv 1zrgVzP linux-by BuckNaked - Debian
Bless me if I can figure this out. The kernel headers were installed, module-assistant appears to see them early on, but later complains that they are not available. The install also complains of package build-essential which is listed as a package on packages.debian.org, but can not be installed either manually or by module-assistant. I am going to investigate this business with build-essby BuckNaked - Debian
root@debian:~# cd /boot root@debian:/boot# ls -l total 27848 drwxr-xr-x 3 root root 4096 Jan 23 17:23 . drwxr-xr-x 22 root root 4096 Jan 23 17:36 .. -rw-r--r-- 1 root root 1991026 Oct 22 01:51 System.map-3.17.0-kirkwood-tld-1 -rw-r--r-- 1 root root 129088 Oct 22 01:51 config-3.17.0-kirkwood-tld-1 drwxr-xr-x 2 root root 4096 Oct 17 19:01 dts -rw-r--r-- 1 root root 6114271by BuckNaked - Debian
The linux headers don't appear to be present and don't appear to be available. Results: root@debian:/etc/apt# cd /boot root@debian:/boot# root@debian:/boot# apt-get install linux-headers-3.17.0-kirkwood-tld-1 Reading package lists... Done Building dependency tree Reading state information... Done E: Unable to locate package linux-headers-3.17.0-kirkwood-tld-1 E: Couldn'by BuckNaked - Debian
Bodhi, thanks for your quick reply. This is what occurred: root@debian:~# apt-get clean root@debian:~# dpkg --clear-avail root@debian:~# rm /var/lib/apt/lists/* --force rm: cannot remove `/var/lib/apt/lists/partial': Is a directory root@debian:~# dpkg --configure -a root@debian:~# module-assistant --verbose --text-mode auto-install xtables-addons-common Updating info about xtableby BuckNaked - Debian
I'm having trouble installing the xtables-addons-common loadable kernel module on my Pogoplug E02. In the past I've always built a custom kernel for my servers and included the iptables support in the kernel, including the necessary extensions (geoip and TARPIT). However, in this case ... due to the patches necessary to make debian run on the Pogoplug ... I'm a bit hesitant toby BuckNaked - Debian
ch, I am no E02 series expert but it sounds like the rescue kernel has been made non-functional somehow. If you have the serial console, I would reload the Doozan rescue kit using the serial console, and then proceed from there using the procedure at the web page above. bodhi can probably give better advice on this situation.by BuckNaked - Debian
No problem. The source kit (as modified by make for the Pogoplug configuration) and the executable will be found in the .tar.bz2 at the following link: http://www.tempfiles.net/download/201308/315874/cgminer.html This is a short-term link and will probably only last a few weeks. Unfortunately the file is too large to attach to this post even after bzipping -9.by BuckNaked - Debian
I will do that. I wrote up the procedure for posterity and it can be seen at http://www.miim.com/thebside/pogoplug/index.html. I will be glad to have any feedback or suggestions on that. Anyone with a Pogoplug V2 and Butterfly Labs ASIC miner who wants to avoid compiling cgminer, let me know and I will be glad to provide a copy of the one I compiled. It was not difficult to compile, justby BuckNaked - Debian
Bodhi, thanks for your help with this problem. I tried the ext2 to ext3 conversion offline several different ways and none worked. After thinking about it a while, I modified the kirkwood.debian-wheezy.sh file to use mkfs.ext3, which was unsuccessful. Eventually I found a solution: From rescue, manually create the /dev/sda1 filesystem using mkfs.ext3 Edit the /kirkwood.debian-wheezy.sh tby BuckNaked - Debian
Will do. I don't have a Linux system or a system drive ready to go ... I shipped my Neoware thin clients out as web servers ... so I'll need to download a live distro to do the ext3 conversion. I tried using tune2fs on the rescue system, and the necessary commands don't seem to be present. It's going on 1 AM here so I'll need to do this tomorrow. I'll update asby BuckNaked - Debian