Re: [Soekris] net5501: FreeBSD ipfw and the elusive 75Mbps throughput -SOLVED
On Jun 18, 2016, at 9:13 AM, Jed Clear <cl...@alum.mit.edu> wrote: On Jun 14, 2016, at 8:50 PM, Jed Clear <cl...@alum.mit.edu> wrote: > On Jun 9, 2016, at 11:01 PM, Andrew Atrens <and...@atrens.ca> wrote: >> On 2016-06-09 8:47 PM, Jed Clear wrote: >>>> Thanks for the replies so far. Looks like I’ll have to wait until >>>> Saturday to test further. Starting with an L2 bridge seems like a good >>>> baseline to try. Although will probably take the easier step of just NAT >>>> w/o rules first. >>> At it's most basic, an l2 bridge can be created using - >>> >>> ifconfig bridge0 create >>> ifconfig bridge0 addm vr0 addm vr1 up >> >> Had an interesting time getting this working. First no “device if_bridge” >> in my kernel (and nanobsd set to not install any kernel modules). Installed >> a new kernel and configured the bridge. But can’t DHCP across the bridge0. >> Finally had to directly attach the laptop to cable modem, let it DHCP and >> then reinstall the net5501 bridge. At that point I was able to download at >> 83. While directly connected to do the DHCP, the same test got 90. But was >> GbE to the cable modem. So I’m thinking 83 is pretty good for 100BASE-T >> interfaces. >> >> The bridge test didn’t come off until now because I’d forgotten a few real >> life things I had to do. But I did do some more thinking and googling >> during the time away. I don’t think I mentioned that I’m still set up to >> do NAT with natd and ipfw divert. Got to thinking that switching in and out >> of the kernel context a few times a packet might not be too good for >> throughput. So next I’m going to see if I can change that over to ipfw >> kernel NAT. Don’t even recall that there was a kernel nat option when I >> first set this up, many, many moons ago. Probably have to add another >> kernel option…. > > Of course it required a new kernel option. In fact it required two. I will > spare you the tale of figuring the second one out. As many have commented on > other boards, ipfw kernel NAT is not well documented. > > But it was worth it. I now get 82 Mbps download through the 5501, with > essentially the same firewall rule set. I did drop dummynet and the inbound > server NAT rules as I no longer have a static IP and I haven’t decided if I’m > going the dynDNS course or sign up for external hosting/VPS/cloud. And I > believe inbound FTP will no longer be an an option as the “punch” dynamic > rules only work with natd. But FTP is no loss. One loose end, polling. I flipped that back on just now and still tested at 83. With the earlier results, it would imply polling and natd is a very band combination for performance. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net5501: FreeBSD ipfw and the elusive 75Mbps throughput -SOLVED
On Jun 14, 2016, at 8:50 PM, Jed Clear <cl...@alum.mit.edu> wrote: On Jun 9, 2016, at 11:01 PM, Andrew Atrens <and...@atrens.ca> wrote: > On 2016-06-09 8:47 PM, Jed Clear wrote: >>> Thanks for the replies so far. Looks like I’ll have to wait until Saturday >>> to test further. Starting with an L2 bridge seems like a good baseline to >>> try. Although will probably take the easier step of just NAT w/o rules >>> first. >> At it's most basic, an l2 bridge can be created using - >> >> ifconfig bridge0 create >> ifconfig bridge0 addm vr0 addm vr1 up > > Had an interesting time getting this working. First no “device if_bridge” in > my kernel (and nanobsd set to not install any kernel modules). Installed a > new kernel and configured the bridge. But can’t DHCP across the bridge0. > Finally had to directly attach the laptop to cable modem, let it DHCP and > then reinstall the net5501 bridge. At that point I was able to download at > 83. While directly connected to do the DHCP, the same test got 90. But was > GbE to the cable modem. So I’m thinking 83 is pretty good for 100BASE-T > interfaces. > > The bridge test didn’t come off until now because I’d forgotten a few real > life things I had to do. But I did do some more thinking and googling during > the time away. I don’t think I mentioned that I’m still set up to do NAT > with natd and ipfw divert. Got to thinking that switching in and out of the > kernel context a few times a packet might not be too good for throughput. So > next I’m going to see if I can change that over to ipfw kernel NAT. Don’t > even recall that there was a kernel nat option when I first set this up, > many, many moons ago. Probably have to add another kernel option…. Of course it required a new kernel option. In fact it required two. I will spare you the tale of figuring the second one out. As many have commented on other boards, ipfw kernel NAT is not well documented. But it was worth it. I now get 82 Mbps download through the 5501, with essentially the same firewall rule set. I did drop dummynet and the inbound server NAT rules as I no longer have a static IP and I haven’t decided if I’m going the dynDNS course or sign up for external hosting/VPS/cloud. And I believe inbound FTP will no longer be an an option as the “punch” dynamic rules only work with natd. But FTP is no loss. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net6501 clock drift
On Jun 18, 2016, at 6:37 AM, Embedding Linuxwrote: > We have a bunch of net6501 boxes around here. The hardware clock suffers > from quite high clock drift (4 to 10 secs a day, depending on the box). > > We are using NTP, so the soft clock is in sync (this is how we > discovered the hw clock drift). But when the boxes reboot, the soft > clock starts from the hw clock, and getting back to sync takes time (up > to several days). You didn’t mention what OS you’re using (although I suppose your e-mail handle is a clue). However it is customary to write the system time to the hardware clock on shutdown, assuming you’re doing a graceful shutdown/reboot. Otherwise look at adding a cron job to do the same thing on a periodic basis. I would also look at doing an initial “ntpd -q -G” (formerly ntpdate) on boot, before starting ntpd normally. Which should make the value of the h/w clock irrelevant, except for early boot logging. And even then, ntpd will log how far the time is adjusted, so you could reconstruct accurate time for the early log values if this is critical. > Is such a big hw clock drift usual on 6501 ? Can’t answer for the 6501 specifically, but PC time of day clock crystals have traditionally had a huge variation in quality. And I don’t think any of them are temperature compensated. But it’s been a long time since I read up on this, so YMMV. Of course being a h/w geek, I added a Garmin GPS puck to my 5501 to keep time. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net5501: FreeBSD ipfw and the elusive 75Mbps throughput
On Jun 9, 2016, at 11:01 PM, Andrew Atrens <and...@atrens.ca> wrote: On 2016-06-09 8:47 PM, Jed Clear wrote: >> Thanks for the replies so far. Looks like I’ll have to wait until Saturday >> to test further. Starting with an L2 bridge seems like a good baseline to >> try. Although will probably take the easier step of just NAT w/o rules >> first. > At it's most basic, an l2 bridge can be created using - > > ifconfig bridge0 create > ifconfig bridge0 addm vr0 addm vr1 up Had an interesting time getting this working. First no “device if_bridge” in my kernel (and nanobsd set to not install any kernel modules). Installed a new kernel and configured the bridge. But can’t DHCP across the bridge0. Finally had to directly attach the laptop to cable modem, let it DHCP and then reinstall the net5501 bridge. At that point I was able to download at 83. While directly connected to do the DHCP, the same test got 90. But was GbE to the cable modem. So I’m thinking 83 is pretty good for 100BASE-T interfaces. The bridge test didn’t come off until now because I’d forgotten a few real life things I had to do. But I did do some more thinking and googling during the time away. I don’t think I mentioned that I’m still set up to do NAT with natd and ipfw divert. Got to thinking that switching in and out of the kernel context a few times a packet might not be too good for throughput. So next I’m going to see if I can change that over to ipfw kernel NAT. Don’t even recall that there was a kernel nat option when I first set this up, many, many moons ago. Probably have to add another kernel option…. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net5501: FreeBSD ipfw and the elusive 75Mbps throughput
> On Jun 9, 2016, at 11:01 PM, Andrew Atrens <and...@atrens.ca> wrote: >> On 2016-06-09 8:47 PM, Jed Clear wrote: >> With the current set up, I ran top during the download. Never got lower >> than 25% idle time on the CPU. ~30% system and and 40+% interrupt. 384M >> (of 512M) Free on memory, so no issue there. So doesn’t seem to be pegging >> the CPU with my full rule set. > That's interesting. The relationship between cpu use and throughput is > pretty linear. You should be able to 'peg' your cpu unless something > (ipfw?) is somehow throttling. You don't have, by chance, 'options > DUMMYNET' configured? My regular rules do have dummy net queues on the uplink, which are now OBE. And I had initially overlooked the bandwidth setting there. In any event, the built in "simple" rule set doesn't I think. > It might also be an tcp-ack-prioritization issue. You did mention that > your uplink speed was kind of crappy. Uplink saturation will affect > downlink speed if tcp acks aren't getting upstream quickly enough. > Maybe ipfw is somehow exacerbating that. 5-6Mbps, crappy only in that $BIG_CABLE_CO disabled 5 of the 8 uplink channels when they "provisioned" my cable modem. > If the l2 bridge thing works and you're able to peg your net5501 cpu > then, notwithstanding vr driver fixes you're probably at the limit. > > BUGS > The vr driver always copies transmit mbuf chains into longword-aligned > buffers prior to transmission in order to pacify the Rhine chips. If > buffers are not aligned correctly, the chip will round the supplied > buffer address and begin DMAing from the wrong location. This buffer > copying impairs transmit performance on slower systems but cannot be > avoided. On faster machines (e.g. a Pentium II), the performance > impact > is much less noticeable. Interesting. You'd think the drivers would align the packets on Rx and I vaguely recall that FreeBSD managed to go zero copy a few major versions ago. Although that could be for straight forwarding. IPFW and NAT might behave differently. Not sure I'd have any control over it. Also recalled that I was running ntpd and named (caching) on the 5501. Turning off ntpd added 5 to the download speed. Turning both off didn't increase that. Didn't think to unplug the GPS which causes a 1 PPS DCD interrupt as well as the 4800 bps serial interrupts. Will give that a try, too. I have the PPS option in the kernel if anyone knows of an interaction. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net5501: FreeBSD ipfw and the elusive 75Mbps throughput
Thanks for the replies so far. Looks like I’ll have to wait until Saturday to test further. Starting with an L2 bridge seems like a good baseline to try. Although will probably take the easier step of just NAT w/o rules first. Switching to pfsense or a 6501 were already further down my option list, so agree with those. Had not considered just a change to pf, but will mull over. Although if I’m going to learn a new syntax, might as well get the pretty interface to go with it (pfsense) assuming I can get the performance. With the current set up, I ran top during the download. Never got lower than 25% idle time on the CPU. ~30% system and and 40+% interrupt. 384M (of 512M) Free on memory, so no issue there. So doesn’t seem to be pegging the CPU with my full rule set. -Jed On Jun 8, 2016, at 8:18 PM, Jed Clear <cl...@alum.mit.edu> wrote: > I just climbed out of the bronze age of home networking (DSL) and now have > "75Mbps service” from $BIG_CABLE_CO (iron age?). Before the DSL was the > bottle neck. Now it appears the 5501 is the bottle neck. My net5501-70 has > long been running nanobsd (FreeBSD 9.3-R) and ipfw as my perimeter > router-firewall-nat. While I’m not expecting 75, especially in the evening, > it’s not even close. Note all the speeds mentioned are download speeds in > Mbps. The upload is much worse, but not bothered by that in this exercise. > > When the cable modem was first brought up, a laptop directly on it pulled 56 > with one of the speed test sites. The cable modem channel power and SNR > don’t look bad. Putting the 5501 in-line dropped the speed to the 30s. Some > googling later and I discover FreeBSD’s polling feature. So I added options > DEVICE_POLLING to the kernel config (HZ was already 1000), baked a new image, > set all the interfaces to polling and … it dropped like a rock to 5 Mbps. > Flipping off polling on the three interfaces brought it back to the 30s. > > I tried the built in “simple” firewall rule set, and that did modestly better > than my, perhaps overly complicated, rule set. It got around 44. I will > work that later. > > Anyway I’m a bit baffled by the negative results when enabling polling. And > any other advice on improving the performance through the 5501 would be > appreciated. I haven’t given up on self help, but need a break from google > for a bit so will appeal to the collective wisdom of soekris-tech. > > Thanks, > > -Jed > > PS: To add insult to injury, I just repeated the directly connected laptop > experiment and clocked over 90. :-( > > ___ > Soekris-tech mailing list > Soekris-tech@lists.soekris.com > http://lists.soekris.com/mailman/listinfo/soekris-tech > ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
[Soekris] net5501: FreeBSD ipfw and the elusive 75Mbps throughput
I just climbed out of the bronze age of home networking (DSL) and now have "75Mbps service” from $BIG_CABLE_CO (iron age?). Before the DSL was the bottle neck. Now it appears the 5501 is the bottle neck. My net5501-70 has long been running nanobsd (FreeBSD 9.3-R) and ipfw as my perimeter router-firewall-nat. While I’m not expecting 75, especially in the evening, it’s not even close. Note all the speeds mentioned are download speeds in Mbps. The upload is much worse, but not bothered by that in this exercise. When the cable modem was first brought up, a laptop directly on it pulled 56 with one of the speed test sites. The cable modem channel power and SNR don’t look bad. Putting the 5501 in-line dropped the speed to the 30s. Some googling later and I discover FreeBSD’s polling feature. So I added options DEVICE_POLLING to the kernel config (HZ was already 1000), baked a new image, set all the interfaces to polling and … it dropped like a rock to 5 Mbps. Flipping off polling on the three interfaces brought it back to the 30s. I tried the built in “simple” firewall rule set, and that did modestly better than my, perhaps overly complicated, rule set. It got around 44. I will work that later. Anyway I’m a bit baffled by the negative results when enabling polling. And any other advice on improving the performance through the 5501 would be appreciated. I haven’t given up on self help, but need a break from google for a bit so will appeal to the collective wisdom of soekris-tech. Thanks, -Jed PS: To add insult to injury, I just repeated the directly connected laptop experiment and clocked over 90. :-( ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Limitations of Net6501 as a network bridge
I haven't explored iptables in bridge/layer 2 mode, but there is no fundamental reason you can't packet sniff or firewall traffic in bridge mode. The traffic has to pass through your kernel. If you were asking about FreeBSD and ipfw, I'd say yes, just look into the layer 2 filter points. So I'd recommend digging into the iptables documentation for something similar. Also not familiar with snort, but with tcpdump, I'd specify the interface and I'm pretty sure it would work as it is capturing at layer 2 (or am I thinking of Wireshark?). -Jed On Jun 25, 2015, at 8:56 PM, Nick Gyurov ngyu...@gmail.com wrote: Bridging works on OSI levels 1 2, IP is a level 4 protocol, so IP services - not so much. I'm not sure what can be done with snort, no experience with it. - Regards, Nick On Fri, Jun 26, 2015 at 6:51 AM, Robin Kipp mli...@robin-kipp.net wrote: Hi all, I just successfully installed Debian 8 on a Net6501. Basically, I’d like to use this device as a transparent network gateway between my computers / network appliances and the router. The reason I’d like to do this is to use an IDS system such as Snort for security purposes and to monitor certain aspects of my personal web traffic. To accomplish this, I’ve now configured a network bridge named br0 and added all of the ethernet interfaces as members. This seems to work pretty well as far as I can tell, I have eth0 hooked up to the router and my other machines connected to the 3 remaining ports. Of course, this means that the Net6501 is pretty much transparent now and does not, for example, show up as a gateway when running a traceroute command to an external host. I generally like this a lot and think it’s quite sensible to pass things like DHCP and NAT on to the router, rather than installing a separate DHCP server on the Net6501 and using iptables to NAT between the Soekris box and the router’s subnet. Since I’m doing this for the very first time, however, I’m wondering whether I’m on the right path, or if such a setup would impose certain limitations that I might not be aware of. For example, can I use a tool like Snort on the Net6501 and analyze all the traffic passing through the Net6501, or could I be facing any issues when doing this on a bridged network? Also, I’m guessing that with this approach, any routing / firewalling of traffic (e.g. by using iptables) is out of the question, right? For example, I would think that it’s probably not possible to use iptables to block traffic originating from certain IP addresses, or to route certain traffic through a configured VPN tunnel, etc… If any of you guys have more experience with network bridges, I’d be very happy if you could answer my questions and / or share your observations with me. Like I said I’m doing this for the first time, so there may well be things that I haven’t even thought of yet! :-) Thanks a lot! Robin ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] com speed changing during boot
Assuming OpenBSD still shares some similarity to FreeBSD, there are three places the speed is set. The one not mentioned so far is the console setting for the getty in /etc/ttys. -Jed On Jun 6, 2015, at 8:20 PM, Nikola (Nick) Gyurov ngyu...@gmail.com wrote: It's 'set ConSpeed=9600'. Case-sensitive. All explained here: https://soekris.com/media/manuals/net6501_manual.pdf - Regards, Nick On Sun, Jun 7, 2015 at 7:56 AM, Jeff Simmons j...@j-simmons.net wrote: On Saturday, June 06, 2015 04:20:28 pm Christian Weisgerber wrote: On 2015-06-06, Jeff Simmons j...@j-simmons.net wrote: I have suddenly run into a problem with one of my Soekris net6501 boxes. When I first boot it, the com speed is 19200, but just before the boot prompt it switches to 9600. This is raising problems with upgrading the OS (OpenBSD). Any idea what's wrong, and anything I can do to fix this? That's due to different defaults in the Soekris comBIOS (19200) and the OpenBSD boot loader (9600). If you want to switch the latter to 19200, add stty com0 19200 to /etc/boot.conf. Tried that, doesn't work. It goes at 19200 up until the boot prompt, switches to 9600 after, and then goes to 19200 at the login: prompt. Can't read the /etc/boot.conf file until the HD is mounted. Alternatively, you could set the comBIOS to 9600. How does one do that? I tried several commands after Ctrl-P early in the boot process, but kept getting 'syntax error'. I have about a dozen net6501s running OpenBSD, and this is the first time I've seen this. -- Jeff Simmons j...@j-simmons.net Simmons Consulting - Network Engineering, Administration, Security ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Can't connect to net6501
Assuming you dropped a zero and tried 19,200 and above, I would at least try 9,600. If that fails, it's probably best to try all the lower speeds as well. And try eliminating the crossover. You could also try removing the SSD to see if you get the BIOS to come up. -Jed On Jan 17, 2015, at 9:20 AM, Johannes Graumann johannes_graum...@web.de wrote: Hello, I am trying to start playing with a net6501 whcih I stuck a 120GB SSD into. I connect the box via a crossover serial cable to a iogear Usb to serial RS-323 adapter, moprobe pl2303 on my linux box and when plugging the adapter in, can see /dev/ttyUSB0 comming up via udevadm monitor. Using cutecom, bautrate 1920 and above device, I seem to be able to connect, but no matter what I do (power off/on, reset, etc.) - there's no communication showing up in the cutecom windows ... can anybody spot what I'm doing wrong her? Sincerely, Joh ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Compact Flash lifetime/failure detection?
I am now worried that my Soekris could go the same way. That machine is net5501 running FreeBSD. You can use the nanobsd configuration script to build FreeBSD. It only writes to the CF when you upload a new build of the OS image. And then it only updates alternating halves of the card. It was designed when flash was smaller and writes were a big concern, but it's still FreeBSD. See the FreeBSD Handbook and nanobsd man page for more details. I've been using nanobsd and an old camera CF on my 5501 since I got it. ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Booting FreeDOS from USB bootable key on net6501
First thing that comes to mind is that FreeDOS might not boot headless. If it is designed to, it might reset the console speed. Something to check on before worrying about more intricate possibilities. -Jed On Aug 12, 2014, at 3:58 PM, ML mail mlnos...@yahoo.com wrote: Hello, I would like to boot from my USB bootable key which contains FreeDOS 1.1 on my net6501 in order to upgrade the BIOS of a few SATA hard drives (I have an eSATA connector attached to my net6501). Unfortunately when I issue a boot 81 nothing happens. Any clue what I am doing wrong? or is it simply not possible to boot FreeDOS on a net6501? Regards ML ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] 32 GB SanDisk Ultra hangs net5501
Something basic perhaps, like did the boot sector and loader get properly installed? Although I would expect some complaint from the comBios in that case. Next thing, is you OS image configured for a serial console in all boot stages. I can only speak about FreeBSD, where ISTR three files need touching. -Jed On Jun 10, 2014, at 5:43 AM, Sebastian Reitenbach sebas...@l00-bugdead-prods.de wrote: Hi, I bought a SanDisk Ultra 32GB, in order to upgrade the disk in my net5501, however, with the new CF card plugged in, the box doesn't boot it stops here: comBIOS ver. 1.33 20070103 Copyright (C) 2000-2007 Soekris Engineering. net5501 0512 Mbyte MemoryCPU Geode LX 500 Mhz Pri Mas SanDisk SDCFHSNJC-032G And then just hangs, I waited a couple of minutes before power cycled the box. I tried with all three comBIOS version available from the downloads site [1] More googling, I found some related issues with solutions on that page here [2] However, also setting the flash=secondary only changed the last line starting with Pri Mas to Pri Sla but it still did hang. I tried this CF card on my desktop, and I was able to access it without issue. Other than having the COM speed set to 9600, no other changes to the settings of the box. Any other ideas I could approach, besides trying different CF cards from other vendors? cheers, Sebastian [1] http://soekris.com/downloads.html [2] http://soekris.com/Issue0006/ ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Net6501 error
On Jan 18, 2014, at 5:46 AM, ws8pnil...@sneakemail.com wrote: i have a problem with my net6501-50. It was working well since 2 years but today it didn't boot. I have the power and error leds lit and nothing on the serial console. I have the same problem with my net6501-50. Sometimes it helps to enter the management console (enter +++ over the serial console and wait one second) and just pretend to update the BIOS (download command) but let the zmodem transfer time out. After that just hit the hardware reset switch. Most of the time mine will boot then. It seems not working for me. Did you see something on the console when you type +++. I have nothing on the console (after i power cycle or type +++ on the console). Also when i press the reset button only the 4 network leds flash. And the error led is always lit. Probably a long shot, but if you had set the serial port speed in COMBIOS to other than the default, 19,200, try setting your terminal to 19,200 and see if you see any error messages or can get the +++ to work. Also, if you have another Soekris box, verify that your terminal is working (again a longshot with that error LED on). IIRC, the error LED will stay on if it cannot access any bootable media (somebody correct me if I'm wrong). Try verifying your boot media on another computer, or at least reseating all its connections. Can i do something else ? Of course you can. My personal favorite (at least in nice weather) is to take a walk over to the local establishment and have them pull me a mug of my favorite ale. Do i need to return the motherboard ? Possibly, but one more thing to try is a different power supply. The OEM wall wart supplies that Soekris resells don't have a great track record. The two I've had failed, producing different symptoms. So it's always worth trying a different one when unexplained problems are encountered. If some soekris engineer read this message, perhaps they have already seen this problem. Occasionally you will see a reply from THE Seokris himself. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Net6501 error
On Jan 18, 2014, at 9:52 AM, ws8pnil...@sneakemail.com wrote: Le 18 janv. 2014 à 15:13, Jed Clear jclear-at-speakeasy.net |Soekris/Forward to (FAC)| dy7lscn...@sneakemail.com a écrit : IIRC, the error LED will stay on if it cannot access any bootable media (somebody correct me if I'm wrong). Try verifying your boot media on another computer, or at least reseating all its connections. But i doesn’t see the bios so i think it never go do far. While double failures aren't the norm, it would still be worth your while to confirm your boot media is in working order. Here's a scenario: you get a power spike or surge that damages the serial port because it was connected to a terminal with different grounding and you were booting off an external disk enclosure with yet another ground. Admittedly this is even less likely if you are using an internal USB drive. But the more you eliminate, the better. Also if you're using an internally powered hard drive, it could be causing a problem with the 12V power. Not sure if Soekris used old school RS-232 voltages, but if so, that's another place that using the 12V power. I'd also strip anything added to the board at this point and see if you get the BIOS with just power and the terminal attached. You'd want to try that anyway before sending the board back. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] HOWTO Soekris BIOS Settings From OS??
On Dec 2, 2013, at 9:31 AM, Hendrickson, Kenneth khend...@harris.com wrote: --- Matthias Flege wrote: ... BIOS reflash from the uManager. I assume that this is a program one can run from an OS which flashes the BIOS. I have a Net5501 on which the console appears to have died. It will boot just fine if the disk is already set up, but there is no serial port console. Neither is there a POST message, or the Have you checked to see that you don't have a serial speed mismatch? I've encountered times where that would result in no characters. And the speed while in the BIOS is not necessarily the same as when the OS is running. And the boot loader could even be at a third speed. Also worth checking with flow control disabled/ignored. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net6501 won't cold boot (will reboot without problems)
Do operational considerations let you swap that msata into another 6501, to see if the problem follows or not? -Jed On Sep 12, 2013, at 1:50 PM, Michael Sierchio ku...@tenebras.com wrote: I have a net6501 that will not boot properly from power-on. It's the only one in my stable of many net6501-70s that will not do so. This is with FreeBSD 9.1, but same behavior is observed with other (older) versions. There is no difference when fastboot is disabled or enabled, so I've disabled it here for brevity. If I press Enter at the boot prompt, it boots (the default works). This same mSATA drive will autoboot on all the other boxes I have. Has anyone run across this before? It doesn't do it every time, but having a box that doesn't reliably boot after a power event is - er - suboptimal. - M schnipp hier POST: 0123456789bcefghipsajklnopqr,,,tvwxy comBIOS ver. 1.41c 20121115 Copyright (C) 2000-2011 Soekris Engineering. net6501 2048 Mbyte MemoryCPU Atom E6xx 1600 Mhz SATA AHCI BIOS ver. 0.61 20121115 Copyright (C) 2003-2011 Intel Corporation Controller Bus#02, Device#06, Function#00: 02 Ports, 01 Devices Port-00: Hard Disk, INTEL SSDMAEXC020G3 Port-01: No device detected Soekris USB Expansion ROM ver. 1.01 20111203 Initializing Intel(R) Boot Agent GE v1.3.72 PXE 2.1 Build 089 (WfM 2.0) Seconds to automatic boot. Press Ctrl-P for entering Monitor. - FreeBSD/x86 boot Default: 0:ad(0p2)/boot/loader boot: ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] help in booting FreeBSD 9
You should also review the FreeBSD handbook. There was a section on serial consoles. There may be new steps for version 9 that I haven't encountered as I'm sticking with 8.x for a bit. And you can double check the ones I did recall. You could also check the logs on your PXE and TFTP server(s) to see if all the expected files were requested and transferred successfully. -Jed On Aug 8, 2013, at 3:16 AM, Igalson Jacek - Korpo TP jacek.igal...@orange.com wrote: Hi Chris, Thank you for your suggestions concerning comconsole and speed. I set baudrate 9600 consitently in Soekris and server according to loppefaaret advice on this mailinglist. I also set boot_serial=-h according to hint by Jed Clear. The last status of my booting is: boot f0 NSC DP83815/DP83816 Fast Ethernet UNDI, v1.03 Copyright (C) 2002, 2003 National Semiconductor Corporation All rights reserved. Pre-boot eXecution Environment PXE-2.0 (build 082) Copyright (C) 1997-2000 Intel Corporation CLIENT MAC ADDR: 00 00 24 CF D8 F8. DHCP.|CLIENT IP: 192.168.1.45 MASK: 255.255.255.0 DHCP IP: 192.168.1.10 GATEWAY IP: 192.168.1.10 TFTP. TFTP.\|PXE Loader 1.00 Building the boot loader arguments Relocating the loader and the BTX Starting the BTX loader C o n s o l e s : i n t e r n a l v i d e o / k e y b o a r d B I O S d r i v e C : i s d i s k 0 P X E v e r s i o n 2 . 1 , r e a l m o d e e n t r y p o i n t @ 9 e 6 4 : 0 0 f 6 B I O S 6 3 9 k B / 1 3 0 0 4 8 k B a v a i l a b l e m e m o r y F r e e B S D / x 8 6 b o o t s t r a p l o a d e r , R e v i s i o n 1 . 1 ( r o o t @ o b r i a n . c s e . b u f f a l o . e d u , T u e J a n 3 0 6 : 4 0 : 0 1 U T C 2 0 1 2 ) p x e _ o p e n : s e r v e r a d d r : 1 9 2 . 1 6 8 . 1 . 1 0 p x e _ o p e n : s e r v e r p a t h : / u s r / p u b l i c / f r e e b s d 9 0 / p x e _ o p e n : g a t e w a y i p : 1 9 2 . 1 6 8 . 1 . 1 0 | / - \ | / end process hangs. My concern is about the message: Consoles: video/keyboard. Is it OK? Last message is from bootstrap loader: pxe_open:gateway ip: 192.168.1.10 192.168.1.10 is my file server (TFTP and NFS server) Characters | / - \ | / suggest that something is loaded to Soekris, but I may not see it. Could someone interpret the messages and explain what is the stage of installation? Regards, Jacek -Wiadomość oryginalna- Od: Christopher Hilton [mailto:ch...@vindaloo.com] Wysłano: 6 sierpnia 2013 17:42 Do: Igalson Jacek - Korpo TP DW: soekris-tech@lists.soekris.com Temat: Re: [Soekris] help in booting FreeBSD 9 On Aug 5, 2013, at 5:14 AM, Igalson Jacek - Korpo TP jacek.igal...@orange.com wrote: Hello, I am new to the Soekris. I try to install FreeBSD on Soekris 4801-48 with CF memory (8GB). I followed excellent guide for PXE booting FreeBSD 9.x: http://freebsd.so14k.com/freebsd9_pxe.shtml To communicate with Soekris I use terminal Minicom (19200baud and 8N1) on PC with FreeBSD 9.0. I got the following output on terminal: boot f0 Looks like you are getting caught out by a couple of issues with booting FreeBSD on a soekris. 1. The FreeBSD pxeboot infrastructure assumes you are booting onto a machine with a VGA card and keyboard. Once the kernel loads it transfers over to the syscons driver and you can't see the output anymore 2. The Soekris serial port defaults to 19200, 8n1 and the FreeBSD serial console assumes 9600, 8n1. In some of your instructions you'll find a section that has you writing modifying the file '/boot/loader.conf' to have the line: console=comconsole in it. That specifies that the kernel should use the serial port as a console. With just that you should expect to see the kernel boot and then the output will turn to hash because the serial speed is wrong. At this point you could drop out of minicom and restart it with a speed of 9600 baud. Alternatively you can add the line: comconsole_speed=19200 That will keep the console speed at the default for the soekris. There are lots of instructions for pxebooting a soekris into FreeBSD 7, 8, and 9 on the net and there are differences in all of them because the different versions of FreeBSD have had various hangups. -- Chris ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] ODP: help in booting FreeBSD 9
On Aug 6, 2013, at 8:30 AM, Igalson Jacek - Korpo TP jacek.igal...@orange.com wrote: Hi Jed, Thank you for your interest and advice. Concerning serial speed: I switched comBIOS parameter: ConSpeed=9600 and ajusted Minicom in my PC accordingly. Problem with -h: I set boot_serial=-h in file /usr/public/freebsd../boot/default/loader.conf Is this what you had in mind? It's been a while since I set mine up, but that seems reasonable. In results, Minicom captured following message: boot f0 NSC DP83815/DP83816 Fast Ethernet UNDI, v1.03 Copyright (C) 2002, 2003 National Semiconductor Corporation All rights reserved. Pre-boot eXecution Environment PXE-2.0 (build 082) Copyright (C) 1997-2000 Intel Corporation CLIENT MAC ADDR: 00 00 24 CF D8 F8. DHCP.|CLIENT IP: 192.168.1.45 MASK: 255.255.255.0 DHCP IP: 192.168.1.10 GATEWAY IP: 192.168.1.10 TFTP. TFTP.\\PXE Loader 1.00 Building the boot loader arguments Relocating the loader and the BTX Starting the BTX loader C o n s o l e s : i n t e r n a l v i d e o / k e y b o a r dB I O S d r i v e C : i s d i s k 0 P X E v e r s i o n 2 . 1 , r e a l m o d e e n t r y p o i n t @ 9 e 6 4 : 0 0 f 6 B I O S 6 3 9 k B / 1 3 0 0 4 8 k B a v a i l a b l e m e m o r y F r e e B S D / x 8 6 b o o t s t r a p l o a d e r , R e v i s i o n 1 . 1 ( r o o t @ f a r r e l l . c s e . b u f f a l o . e d u , T u e J a n 3 0 6 : 5 1 : 4 9 U T C 2 0 1 2 ) p x e _ o p e n : s e r v e r a d d r : 1 9 2 . 1 6 8 . 1 . 1 0 p x e _ o p e n : s e r v e r p a t h : / u s r / p u b l i c / f r e e b s d 9 0 / p x e _ o p e n : g a t e w a y i p : 1 9 2 . 1 6 8 . 1 . 1 0 | / - \ | / Can you guess where we are and what is missing? I think your missing something in your PXE setup, but since I don't use PXE myself, I'm going to defer to others on the list. Regards, Jacek -Wiadomość oryginalna- Od: Jed Clear [mailto:jcl...@speakeasy.net] Wysłano: 5 sierpnia 2013 12:57 Do: Igalson Jacek - Korpo TP Temat: Re: [Soekris] help in booting FreeBSD 9 There are three places you need to adjust to deal with the 19200 serial console in FreeBSD, IIRC. You need to tell the loader to use 19200. You need to pass -h to the kernel. And finally you need to change the speed of the getty. Some of us find it more expedient to change the BIOS speed to 9600 so we don't have to adjust the other two. Still need the -h. -Jed On Aug 5, 2013, at 5:14 AM, Igalson Jacek - Korpo TP jacek.igal...@orange.com wrote: Hello, I am new to the Soekris. I try to install FreeBSD on Soekris 4801-48 with CF memory (8GB). I followed excellent guide for PXE booting FreeBSD 9.x: http://freebsd.so14k.com/freebsd9_pxe.shtml To communicate with Soekris I use terminal Minicom (19200baud and 8N1) on PC with FreeBSD 9.0. I got the following output on terminal: boot f0 aC DP83815/DP83816 Fast Ethernet UNDI, v1.03 npyright (C) 2002, 2003 National Semiconductor Corporation rl rights reserved. e te-boot eXecution Environment PXE-2.0 (build 082) npyright (C) 1997-2000 Intel Corporation i :IENT MAC ADDR: 00 00 24 CF D8 F8 sIENT IP: 192.168.1.45 MASK: 255.255.255.0 DHCP IP: 192.168.1.10 eTEWAY IP: 192.168.1.10 lE Loader 1.00 o silding the boot loader arguments nlocating the loader and the BTX oarting the BTX loader C I see that in the end of printout cursor jumps to the line just next to boot f0 runs forwards and backwards several times than stops and terminal hangs. Port Eth0 shows activity, LEDs are on but there is no answer for ping and ssh. Is something wrong with terminal parameters? I use 19200baud 8N1 and VT102 or ANSI, is that ok ? Or boot loader BTX is working badly. I would appreciate your advice and help. Jacek ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Problem with net6501
On Jan 19, 2013, at 8:18 PM, Christian Lavoie christian.lav...@gmail.com wrote: Ok, quick update -- managed to make it boot after removing the lan1841 that I ordered with it (OpenBSD now installed and working like a charm). So, as far as I can tell: with the lan1841 the main CPU refuses to boot, and uManager doesn't see the Flash BIOS (df reports zeroes everywhere I checked by hand). Without the lan1841, everything works fine (except I'm short 4 network ports, obviously). It seems a tight fit to plug it in, is the rackmount case known to be 1-2mm too short? If you haven't already, you might want to see what happens if you remove the 6501 from the case and try it with the lan1841 on the bench. It's possible, if there was a tight fit, that something was distorted enough to short out. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net6501 GPIO
From the description, they would appear to be writeable addresses that would allow you alter select outputs without knowing the current state of the output. So writing an 0x0080 to 0688 would clear bit 7 to 0 and leave the others in their current state. Writing 0x0808 to 068c would set bits 11 and 3 to 1 without changing any other. That's how I read it based on what you just posted, so YMMV. -Jed On Jan 10, 2013, at 8:03 PM, Matt Dainty m...@bodgit-n-scarper.com wrote: I've been writing a driver for the GPIO as implemented by the FPGA on the net6501 and I was wondering if someone (Soren?) could clarify the purpose of some of the I/O ports. From the archive, I have the following information regarding the GPIO: I/O Port 0680 bit 15..0 gpio input 0684 bit 15..0 gpio output 0688 bit 15..0 gpio reset bit when '1' 068C bit 15..0 gpio set bit when '1' 0690 bit 15..0 gpio dir '0'=input, '1'=output The first, second and fifth I/O ports seem fairly obvious; read state from the first, write state to the second in the case of output pins and change the direction with the fifth but I can't work out what the third and fourth I/O ports are used for, they always seem to read as 0. Could someone explain how they're supposed to be used? Matt -- Never rub another man's rhubarb. ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Bridging OpenBSD
On Apr 25, 2012, at 2:07 PM, Lars Noodén wrote: I am connecting four hosts to the Internet via a net5501 running OpenBSD. I would like also to be able to ping and ssh between these hosts directly rather than having each on a separate subnet. Where should I be looking to set up such capabilities? +-net5501--+ | | host1 ---+--vr0--? | | | host2 ---+--vr1--? | | ?--re0--+isp dhcp host3 ---+--vr2--? | | | host4 ---+--vr3--? | | | +--+ My first guess was to add vr0 - vr3 to a bridge and try to assign an IP number to the bridge, but that does not work. I'll assume you've seen http://www.openbsd.org/faq/faq6.html#Bridge Can you expand on does not work? Do you mean host1-4 can't ping each other? Or they can't ping the IP on the net5501 bridge interface? Or they can't reach the Internet through the net5501 (assuming that is the ultimate goal)? If hosts1-4 can't ping each other, do you still have /etc/hostname.vr0-3 with the up? Can you post the output of ifconfig -a? Any network related log entries from a reboot? -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net6501: are the E6xx LVDS pins available as pads anywhere?
On Mar 1, 2012, at 8:28 PM, Andrew andrew_soek...@16paws.com wrote: Hi, I've been looking into upgrading my aging firewall from a pentium-m running FBSD 4+ to something more recent. Since I've stumbled over the net6501, I have tried researching it as much as possible. My firewall is also my GPS strat-1 NTP server, UPS monitor and DHCP server, among other services. Because I use a KVM to switch VGA between servers, I am looking at a simple VGA solution for the net6501. Options using the PCIe x1 are constrained by Intel's crippling of the bus to version 1.0a - the AMD FirePro 2270 x1 is PCIe 2.1, the nVidia Quadro series is PCIe 2.0, and HIS ATI Radeon HD5450 is 2.0 (as well as doubling the power requirements). The onboard Atom E6xx does include Intel's third tier graphics on the die, with separate pipes to LVDS and SDVO outputs. Converting LVDS to VGA is a simple $30 problem. Does anyone know if the LVDS signal are exposed as pads or are the balls not connected? Additionally, looking at the EG20T datasheet, it appears that only UART0 includes hardware handshake signals (8-wire interface). The other 3 UARTS only bring out Rx Tx signals (2-wire interface). This gets to be a problem implementing a PPS NTP kernel which requires the handshake lines, unless the console serial port is not UART0. Or the internal video is available (and, obviously, the BIOS supports it, ugh). I've had no problems with a Garmin GPS 18 LVC feeding a PPS FreeBSD kernel via the internal serial port's DCD pin on my 5501. Based on Soren's response, it sounds like the 6501 is similarly endowed. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] vr0 using OpenBSD stops responding.
On Dec 10, 2011, at 7:50 AM, Frank Schuhmann wrote: Because you wrote that the Win XP was not producing the problem, but Win 7 should do it, it is perhaps the TCP window scaling, as declared in the RFC 1323. In some similar other cases it was sometimes fixing the problem (pfsense) and in sometimes also not, but to quick test it out the spended time will to be of value. On the BSD side perhaps you try to turn of the scaling, I´m not an OpenBSD professional but it must be turn out by setting up a shell order likes sysctl -w net.inet.tcp.rfc1323=0 If not or it must be typed in, in other direction it will be super that perhaps an OpenBSD familiar list member would correct this please. RFC1323 affects the end to end TCP connection. Unless there is a proxy running on the 5501, its RFC1323 settings won't change things. But I agree with this, and the previous poster, that a likely cause is Win7 using RFC1323 and other TCP optimizations (e.g. SACK) for more efficient transfers, so the 5501 is working harder. However another thing occurs to me. I wonder how Path MTU is being dealt with in this scenario. I have far too much experience[1] with various tunneling methods (e.g. VPN) that do not correctly support Path MTU Discovery. We ended up with the ugly, but effective, kludge of artificially lowering the MTUs on the server and client network stacks to compensate. And about half the time when we deploy something new, it breaks until we remember to change the default MTU setting. Perhaps these new Windows 7 boxes don't have that applied, so you're getting a lot of fragmenting at the 5501, and possibly reassembly (although often this is left to the receiving host to do). Seems like a long shot as the usual symptoms for us is no bulk traffic, but YMMV. -Jed [1] The painful sort. ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Official net6501 bios update
On Dec 5, 2011, at 3:03 PM, Jim Pingle wrote: I did make one change. Because that image does not have the serial console enabled I mounted the image on another FreeBSD box and activated the serial console in /boot/loader.conf.local like so: boot_multicons=YES boot_serial=YES Did you create a /boot.config file with -h in it? In fact that's all I have done, I don't have the two settings above in my loader.conf file. Of course I run my console at 9600 so didn't need those. To run at 19200, you also need comconsole_speed=19200 in your loader.conf[.local] file. You also need to edit /etc/ttys to keep the 19200 speed. I found the path of least resistance was to change the comBIOS speed setting to 9600 once, so I could use the FreeBSD defaults w/o having to remember to tweak each build in multiple places. I'm running NanoBSD 8.2R-p2 on a net5501. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Updating net6501 on Mac OS problem
On Dec 3, 2011, at 12:21 PM, Tom Weustink wrote: I am using Zterm on Mac OS X (10.7.2) since afaik that is the only terminal program able to send data files. I tried sending using XMODEM as described on the wiki page for updating, and the file seems to arrive alright. But when I issue the flashupdate command, all I get is Error, Incorrect signature of BIOS image. I first downloaded the BIOS file with Safari, and second with a terminal command (wget alias to curl -O). No luck on both sadly. Anything that I am missing here? Hope to get this box booting from USB as it's been laying around for almost 2 months now. At a guess, are you working at the commBIOS Monitor prompt?Updating via commBIOS is broken until you get this update installed. Catch-22, so you need to use the lower level uManager, as described in Soren's e-mails to the list. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] 5501-70 AC adapter stops working
On Tue Nov 15 10:18 , Jay Em Cee sent: I've experienced this a couple of times now. The first time was quite perplexing, and it was only on a lark that I tried a spare AC adapter that I had. I thought the whole unit had gone tango uniform. I don't have skills to disassemble the AC adapter and figure out where things have gone wrong inside. The mode of failure is that the unit won't boot; it doesn't even get far enough in the boot process to start writing output to the serial port, so I get no indication whatsoever as to what's going on. The Power, Error, and Disk LED lights light up, with the disk LED blinking rapidly. There is a faint clicking sound from inside, almost as if the adapter isn't supplying enough juice for the disk to spin up (using a Hitachi 7200 RPM, 5V, 1.1A 100GB disk). I am using Soekris-supplied SMP Technology, Inc.-branded 12V, 1.5A AC adapters, FWIW. I'm asking about this now because I'm out of spare adapters, and I need to order more. I'm wondering if I have a bad batch or if there's something I have mis-spec'd when I built these out. Any thoughts appreciated, If you search the mailing list archives, you'll find you're not the first. I don't recall if I posted about it myself, but I've gone through two. The first time Soekris replaced it immediately, the second time I went with another source. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
[Soekris] NanoBSD Reboot failure
I'm running NanoBSD (FreeBSD 8.2R) on a net5501. It generally runs great, but every time I go to reboot it, I get an error and it doesn't reboot. Same error for halt, reboot, and shutdown. net5501# uname -a FreeBSD net5501 8.2-RELEASE-p2 FreeBSD 8.2-RELEASE-p2 #0: Sat Jul 16 12:26:17 EDT 2011 r...@fbsdam3.my.domain:/usr/obj/nanobsd.custom/i386/usr/src/sys/JEDWALL i386 net5501# halt /sbin/halt: Input/output error. net5501# shutdown now /sbin/shutdown: Input/output error. net5501# reboot /sbin/reboot: Input/output error. net5501# On the console there is a bit more info: login: vnode_pager_getpages: I/O read error vnode_pager_getpages: I/O read error vnode_pager_getpages: I/O read error Fortunately I can reach out and touch it and give it a hard boot when necessary (new builds, generally), but I'd like to fix it. The NanoBSD configuration file I use is included below. -Jed NANO_NAME=custom NANO_SRC=/usr/src NANO_KERNEL=JEDWALL NANO_ARCH=i386 NANO_IMAGES=2 JEDWALL_DIR=/home/clear/jedwall CONF_BUILD=' NO_KLDLOAD=YES NO_NETGRAPH=YES NO_PAM=YES ' CONF_INSTALL=' NO_ACPI=YES NO_BLUETOOTH=YES NO_CVS=YES NO_FORTRAN=YES NO_HTML=YES NO_LPR=YES NO_SENDMAIL=YES NO_SHAREDOCS=YES NO_EXAMPLES=YES NO_CALENDAR=YES NO_MISC=YES ' #NO_INSTALLLIB=YES #NO_MAN=YES #NO_SHARE=YES CONF_WORLD=' NO_MODULES=YES NO_KERBEROS=YES NO_GAMES=YES NO_RESCUE=YES NO_LOCALES=YES NO_SYSCONS=YES NO_INFO=YES ' # FlashDevice SanDisk 1G cust_nobeastie() ( touch ${NANO_WORLDDIR}/boot/loader.conf echo beastie_disable=\YES\ ${NANO_WORLDDIR}/boot/loader.conf ) install_packages () ( mkdir -p ${NANO_WORLDDIR}/packages cp ${JEDWALL_DIR}/pkg/* ${NANO_WORLDDIR}/packages chroot ${NANO_WORLDDIR} sh -c 'cd packages; pkg_add -v *;cd ..;' rm -rf ${NANO_WORLDDIR}/packages rm -f ${NANO_WORLDDIR}/usr/local/etc/thttpd.conf ln -s /etc/www/thttpd.conf ${NANO_WORLDDIR}/usr/local/etc/thttpd.conf ) install_jedwall_files () ( cp -Rp ${JEDWALL_DIR}/Files/etc/ssh ${NANO_WORLDDIR}/etc/ssh cp -RL ${JEDWALL_DIR}/Files/* ${NANO_WORLDDIR} cp /etc/master.passwd ${NANO_WORLDDIR}/etc pwd_mkdb -p -d ${NANO_WORLDDIR}/etc ${NANO_WORLDDIR}/etc/master.passwd ln -s /var/named/etc/namedb ${NANO_WORLDDIR}/etc/namedb ) customize_cmd cust_comconsole customize_cmd cust_install_files #customize_cmd cust_allow_ssh_root customize_cmd cust_nobeastie customize_cmd install_packages customize_cmd install_jedwall_files ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] FreeBSD - Quick help needed...
In some sort of cosmic coincidence I'm literally replacing half a 8.1 GEOM Mirror right now (it's running a disk verification from the controller now). Luckily for me I caught it earlier than you, when only one of the drives was coughing up hard read errors. In the interim I was able to take the parition(s) with the bad reads out of their mirrors, but that's not going to help you any, sorry. In the past I think I kept a drive doing by using dd to just force a write onto that sector, of course that clobbered the file there. In that case I usually had a file name so I would rename the offending file to filename.bad to keep the sector from being unused and then restore the file from backup. Be very careful that you account for sector numbers being offset from partition and slice boundaries if you try this method. And this was probably back in FreeBSD 4.x or 6.x, so no warranties and YMMV. -Jed On May 31, 2011, at 6:20 PM, Soren Kristensen wrote: Hi, Our main server have disk problems, and I need a little help It's running FreeBSD 7.2, with a GEOM mirror in degraded status, one disk is already dead and the other have 2 hard read errors I have tried fsck several times, but it keep insisting on running fsck, refusing fixing those 2 hard read errors. I have new disks in the mail, but this is the main soekris server so I need to get it up and running asap. Any tips on how to get it up and running ? Is it possible to maybe just force the file system clean ? Please copy to so...@soekris.dk as I currently can't read soekris.com mail, that's via the main server And yes, I have new disks in the mail And yes, the old disk was enterprise class drives. Best Regards, Soren Kristensen CEO Chief Engineer Soekris Engineering, Inc. ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Pyamyd login fails
On Apr 20, 2011, at 10:14 AM, l.rine...@movimatica.com wrote: After i boot Pyramid Metrix Pebble/Linux 0.9 in my Soekris net4801 from compact flash, the system prompts: pyramid login: i enter root but the reply is: pyramid login: root Metrix Pebble/Linux 0.9 pyramid ttyS0 pyramid login: and so on, and i am not able to enter in the system. Can you help me ? You'll probably get better results asking this in a Pyramid or Linux forum, but I'll give it a shot anyway. Note that I use FreeBSD, but the Unix login process is fairly similar. YMMV. Something is missing that the login needs to authenticate root and successfully start a shell. Given that it never prompts for the password, I'd check /etc/passwd and all it's adjuncts which vary by OS (some use /etc/shadow, /etc/master.passwd). Make sure the root entry isn't garbled. You could also be missing the root shell, but I don't think you're getting that far. Or the permissions are wrong on any of these. Hmm, maybe even the login executable if it's getty that supplies the prompt. Not quite sure of the division of labor there. If you have Pyramid running on something else try man login for clues and take note of the files listed toward the end. Not sure if Linux uses this, but check /etc/ttys and make sure the console line is marked secure, otherwise direct root login isn't permitted. Try booting to single user, or mount the CF on another system, to check all this. Good luck, -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] *BSD vs Linux
On Apr 13, 2011, at 6:06 PM, Ken Hornstein wrote: I'm really curious why those that chose BSDs chose them? I'm a BSD user from waaay back; I cut my teeth on SunOS 4 in college, and I ran a actual BSD 4.3 system on a derelict VAX 11/750 just for the pure hell of it. I have a similar story to Kens. It all started when a busy coworker handed me some QIC cartridges with SunOS 4.0.1 on them and asked if I wanted to learn Unix. We needed to get a big 9U Sun SPARC VME board working and since I couldn't do anything with the VME cards I was designing until it was loaded, I got the job. What was the GUI back then ... oh, yeah, SunView! But I digress. Anyway SunOS 4 was very BSD, so that's what's seemed right to me ever since. I was around when 386BSD was forming, Not much later, when I wanted a Unix for home on a 386 box, I somehow stumbled on Walnut Creek and FreeBSD 2.0 (.7 I think) CDs. I was also impressed that WC ate their own dog food and ran their busy FTP site on FreeBSD. Pretty much a bullet proof network stack from BSD. Even ATT and Microsoft thought so... but I digress again. This was early '95, or at least it says Jan. '95 on the jewel case. I don't recall if I'd even heard of Lunix back then. When I built my third tower system, I planned to turn the prior one (486DX/66) into a firewall for my new always on DSL line, running OpenBSD because of its hard earned security reputation. But I ran into trouble with the driver for one of the used NICs I'd picked up at a computer show. Even though Theo responded personally to my bug report, I couldn't wait for a fix, so that got FreeBSD as well. When eventually replacing failed parts on that tower became less like fun and more like work, I decided on something smaller and quieter, which led to a net5501 and the discovery of NanoBSD. I hadn't known it, but I had been installing its build tools for years as it's part of the FreeBSD base (thanks Poul). I can echo Ken's comment about cleaner. The BSDs are designed as complete OSs. Each project does their own kernel AND userland (helping themselves to the best of the other projects' code, of course, but more on BSD License in a moment. You can compile a very capable OS solely from each projects code base. None of the *BSDs are distros as they aren't distributing outside code in the base. Of course they do include a lot of optional third party software. The Ports system is just amazing, but no longer that unique to the BSD camp. And if you add the Linux compatibility to the kernel, you can run almost any Linux binary as well. And this all adds up to the control that allows an orderly directory hierarchy. I'll defer to Ken on kernel building being cleaner as I haven't built a Linux kernel. Obviously there is a big difference in the licensing between BSD and GPL. Personally it doesn't make much practical difference to me as I don't write a whole lot of code, so is outside the scope for my response to the original question. But it does matter to many. There's a reason that my second favorite OS, MacOSX, has a FreeBSD middle. And of course Microsoft's original IP stack for Windows was also BSD code. And of course the most important reason for any OS choice, mascots. Beastie is cuter than Tux! OK, I've wasted enough of your bandwidth. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] multiport serial card recommendations for a net5501?
On Feb 27, 2011, at 3:31 PM, Alfred Landrum wrote: If you have good or bad experiences with a multiport serial card, please let me know. I'd like to find an 8 port serial card to use with a net5501 in order to create a linux-based serial access terminal server. I've found several vendors online, but not all specify whether they will support 3.3v. I saw the 1u soekris solution on the antenna-factory.com website uses a sunix ser5066a, but my search attempts didn't locate anyone selling that card in the US. This seems to meet your specifications http://www.cooldrives.com/8sepcilowprc.html It has the same 16C950 UART emulation as the Sunix, claims 3.3V support and has a FL vendor. Note: I'm not vouching one way or another for the maker or vendor since I don't have any experience with either, just pointing it out. On the other hand, if you can't even get a serial port card design right, you're not going to be in the hardware business long. Anyway caveat emptor. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net5501 bricked, need insight on what happened
My recollection is that the error LED for for a few seconds during cold boot is normal. If it stays on you have an issue. My first thought is that you have a speed mismatch on the serial line. Perhaps it didn't stay the same speed after the reboot, or you restarted the terminal emulation and it defaulted back to 9600. That's one reason I always go into CommBIOS first thing and set it to 9600, which is the default for the rest of the world. Many less problems that way. After that, power supplies are a frequent issue, as the other poster mentioned. I'm on my third. Soekris replaced the first one under warranty and when it went, I bought something completely different. If you boot off the CF or HDD can you ping it on the network? Does the HDD even act like it's booting? -Jed On Feb 18, 2011, at 11:37 AM, Andrea Gronchi agron...@smrobotica.it wrote: Hi, I've been experimenting with a net5501 board for some days. The board run Linux (2.6.36) and worked quite nicely botting from PXE+NFS and CF. This same board also spent some time running as a test firewall appliance, without apparent problems. After a clean hard reboot (correct system shutdown followed by power supply disconnection), the board stopped working: - no POST with CF inserted, - no POST with SATA drive connected. - no POST even with a clean, unconnected board. Other than the power led on, the only recognizable sign of activity is the error led, which happens to stay on for the first 6-8 seconds after powering on. I'd like to have some insight in this specific condition, possible fixes and what is causing this. Thank you in advance. Andrea. Andrea Gronchi agron...@smrobotica.it ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] 4G CF errors
On Jan 25, 2011, at 9:41 AM, Michael Smith wrote: Mike Tancsa wrote: ad0: FAILURE - READ_DMA status=51READY,DSC,ERROR error=10NID_NOT_FOUND LBA=7813103 # ls -l _.disk.full -rw-r--r-- 1 root wheel - 972776448 Jan 10 17:11 _.disk.full dd if=_.disk.full ibs=64k obs=64k of=/dev/da0 NANO_MEDIASIZE=190 I'm not familiar with NanoBSD, but does this mean it's a 190 x 512 byte image? Do you know why the kernel would be trying to read from sector 7813103 if the image is only 190 sectors? The comment in the actual nanobsd.sh script says NANO_MEDIASIZE is a count of 512 byte sectors. Note that is the size used to create the _.disk.full file which actually contains three BSD partitions on three different BSD slices (aka DOS partitions), a small partition to write config changes to and then the rest divided into half so you can ping pong between two versions of the OS by loading just the _.disk.image file using one of the update# scripts. So the current value leaves a lot of uninitialized sectors at the end of the CF after the dd, which probably aren't accounted for in the DOS partition table. Also I think certain parts of the GEOM driver check the last sector of a drive for configuration info? Maybe you need to change MEDIASIZE to match the device so those bits of the CF are properly initialized by the dd. At least in 8.1 you can add the line FlashDevice sandisk 4G to your nanobsd.conf file and omit the NANO_MEDIASIZE value. You might want to try that first. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] FreeBSD 8.1: kernel config (kernel: stray irq7)
Another net5501 data point: cl...@net5501:~ vmstat -i interrupt total rate irq0: clk 1024103514 1000 irq3: uart1 25936365 25 irq4: uart0 579 0 irq5: vr16171126 6 irq8: rtc 131068465127 irq11: vr0 6262568 6 irq14: ata0 1761 0 irq15: ata1 ohci0+ 1 0 Total 1193544379 1165 With almost 12 days of uptime on 8.1R-p2, nanoBSD build on a 1GB CF card. There is a Garmin18 GPS hanging on uart1 sending 1 message per second and a DCD interrupt from the 1PPS line I recall those stray irq7s from before, but apparently they are gone on my system. -Jed On Dec 15, 2010, at 4:39 AM, Martin Johnson wrote: On 08/12/2010 19:30, Martin Johnson wrote: After updating from 7.1 to 8.1, I'm seeing kernel: stray irq7 kernel messages in the syslogs on three NET5501's. ... The box is a NET5501-70 with a SATA laptop drive, and no extra cards. So I start to think there's something in my kernel config that doesn't suit the 5501 hardware. I have now trimmed down my kernel config for FreeBSD 8.1. If anyone wants it, it can be downloaded from http://net42.co.uk/downloads/NET5501 Everything seems to work, but top is still reporting up to 10% of CPU time being taken by interrupts, even when the system is idle: CPU: 0.0% user, 0.0% nice, 1.9% system, 9.3% interrupt, 88.7% idle After each boot, the log shows kernel: stray irq7 and later kernel: too many stray irq 7's: not logging anymore, as if there's something on the Soekris board generating IRQs that we aren't handling. I believe IRQ 7 on FreeBSD kernel reports is a placeholder for missed interrupts. Just looking at interrupt counters: # vmstat -i interrupt total rate irq0: clk 158121051 1000 irq3: uart1 2221 0 irq4: uart0 375 0 irq7: 6 0 stray irq7 6 0 irq8: rtc 161876641 1023 irq11: vr0680731 4 irq14: ata0 436168 2 irq15: ata1 ohci0+ 94943420600 Total 416060619 2631 Could the stats for irq0 irq8 reflect a clock that's ticking a bit too quickly for a relatively slow CPU? I don't have anything on ATA1 (just a SATA laptop hard drive on /dev/ad0). But I do have a USB multi-port RS232 device, though as far as I know it's not hit very hard: it just plugs into various device consoles, plus a weather receiver that receives brief bursts of traffic at 19200 baud every 20 seconds or so. If I run vmstat -i on a different NET5501 (without the USB device), then I get: # vmstat -i interrupt total rate irq0: clk 572865511999 irq3: uart1 4559 0 irq4: uart0 334 0 irq8: rtc 73317680127 irq11: vr0443655 0 irq14: ata0 617127 1 irq15: ata1 ohci0+ 1 0 Total 647248867 1129 Once again the clock rate seems a bit high, but the rate for IRQ15 is zero, presumably because there are no USB devices on that machine. Is CPU_ELAN_XTAL required on the NET5501, or was that only for NET4801 hardware? Thanks for any pointers, - Martin. ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] serial-only boot of x86 Solaris 11 Express?
My not terribly recent experience with Solaris on SPARC was that if there was no Sun keyboard installed, it would automagically use TTYA as the console. Which brings to mind a question, what speed do you have your terminal set for? If it's set to the net5501's default 19200, you might just need to set everything (BIOS and terminal) to 9600 to catch Solaris booting. I wouldn't bet on it being that simple, but certainly worth a try. -Jed On Dec 2, 2010, at 8:53 PM, Malcolm Herbert wrote: I'm trying to install Solaris 11 Express on a Soekris net5501 host which has 256MB of memory and no video card for use as a low-power server[1] - how do I configure grub to tell the kernel to use the serial port as a console? after some hardware issues, I managed to get a 32bit install of Solaris 11 Express installed onto a 2.5 SATA drive using my netbook. After making sure that grub didn't use a graphical splash screen I was able to hook the HDD up to the net5501 board and boot it. The serial console showed the grub screen and was I able to watch it count down, however when grub handed off to the kernel, there was some disk activity (around 20-25sec worth) before the host rebooted again ... all the while without any output to the BIOS video emulator or to the serial port ... This is encouraging - I'm expecting the host to not perform well (if it does at all) due to the extremely tight memory, but I'd still like to see where it is choking and see what else the booting kernel discovers as it comes up, however I'm not sure what to do with grub to get the appropriate logging messages to the kernel ... has anyone tried to get a purely serial-only install of Solaris/OpenSolaris working on x86 hardware? No brainer for SPARC, which is where I'm more familiar ... Regards, Malcolm [1] the host I am intending on using does have the minimum required 512MB but it's in use at the moment -- Malcolm HerbertThis brain intentionally m...@mjch.netleft blank ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
[Soekris] ipfw+natd changes in FreeBSD 8.1
Another gratuitous lesson learned from upgrading my net5501 from FreeBSD 8.0R to 8.1R. May it save others who follow. -Jed If you're using natd with ipfw in FreeBSD 8.0 or earlier and upgrade to FreeBSD 8.1, you may need to modify your ipfw divert rules slightly. The following works up through 8.0, but will fail with no indications in 8.1: add 300 divert natd ip from any to any in via vr1 add 6 divert natd ip from any to any out via vr1 You need to change the ip to ip4 in divert rules. For example: add 300 divert natd ip4 from any to any in via vr1 add 6 divert natd ip4 from any to any out via vr1 And since ip and any are supposedly synonyms, I'd guess you'd need to change any in a divert rule to ip4 as well. I saw no explanation on why this is needed. My guess is natd is choking on non-IPv4 packets now, where it ignored them in the past. Note, I use the cpp preprocessor on my ipfw rules, so your syntax might be slightly different. ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] how to xfer an OS/install from a VirtualBoxVM to a running Soekris, over the network?
On Aug 13, 2010, at 12:42 PM, Ben DJ wrote: On Fri, Aug 13, 2010 at 9:35 AM, Martin Johnson martin.johnson.uk.li...@googlemail.com wrote: Assuming that the image you've made really is bootable on the Soekris, If it's booting in the VM, any particular reason it might not on the Soekris? I would PXEBOOT the Soekris box from a Voyage Linux LiveCD on a laptop over an Ethernet cable, then use 'dd' in Voyage Linux to pull the image from the network (or copy from a USB stick) onto the Soekris HDD. The advantage of Voyage Linux is that it's a LiveCD that you can use to provide the PXEBOOT environment using just an old laptop and an Ethernet cable. So there's nothing to configure, and no changes to your normal network. Didn't know abt Voyage Linux. Since I've already got a VM running FreeBSD, what's the specific advantage of Voyage over that? I fear I'm missing your point. I infer that the Voyage Live CD has a PXE boot server already configured and running. Being a FreeBSD person myself, I'm pretty sure FreeBSD does not have such a server configured by default. You'd probably have to find the port, install and configure it. If you've got a CF card around, you could copy your image to CF, boot the 4801 from CF, use that to dd the image to the HDD. Or if your image uses swap, build a basic NanoBSD image for the CF. You could even keep the CF installed as a backup to the HDD. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net5501 and PCI card limitations
On Jul 6, 2010, at 8:34 PM, Nick Lidakis wrote: As there is no manual for the net5501 on the Soekris site I thought I'd ask here. Are there any limitations to what kind of PCI card one can use with the net5501? In building an appliance like music server, I'd like to experiment with several different PCI sound cards. Some require +12v and -12v voltages, as the cards offer analog out and S/PDIF. Often the 4801 manual is referenced for the 5501. It lists a maximum of 0.3A for +12V and 0.1A for -12V on the PCI connector. There is also a 10W combined limit on the 3.3V and 5V PCI pins, shared with any hard disk. Any gotchas? Many, like the above is for the 4801 and may or may not be the same for the 5501. It also depends heavily on the size of the supply you use to power the 5501 itself. See the notes with the various supplies that Soekris offers. One thing that is certain is that you will need a card compatible with 3.3V PCI signaling. If that doesn't curtail your interest, your best bet is to search this list's archives and the wiki (http://wiki.soekris.info/) for an answer. You might find the answer in discussions of telephony cards that come up occasionally. Or wait a while to see if somebody more useful responds. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Net-booting Debian 5?
[Resend to list] On Tue Mar 2 15:18 , Steven Finnegan sent: I have been attempting to net-boot Debian 5 on either a 5501 or a 4801. I have the pxe environment set up. I have been following instructions here: http://wiki.soekris.info/Installing_Debian_Linux_5.0 including the editing of the configuration files. Are you sure you did the 5. Modify the netboot files a little to work with a serial console. Debian used to come with this but it's gone now. step? Completely? When booting the environment, I get to the point of getting the PXELINUX copyright notice, then everything halts. Any help would be appreciated. Screen capture of boot sequence follows... [snip] debian-installer/i386/boot-screens/vesamenu.c32-- Block #1 is read 4 times w/o acknowledgement, no further net accesses That you hang up in the _VESA_ menu is what makes me think something isn't completely configured for a serial console. If not the netboot bits, then later on in the boot sequence. Disclaimer: But that's pretty much a hunch on my part. I use FreeBSD myself and haven't booted a diskless client in a long while, and that was a sun4m and Solaris 5.8, IIRC. HTH, -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Seagate Barracuda 750G SATA not working on net5501
On Sun Feb 28 9:00 , Henrik /KaarPoSoft sent: Dear all, I have a net5501, which seems to work with a Kingston 64GB SSDnowV+ (SATA) and Seagate Barracuda 7200.7 (3.5 160GB) SATA. However, I can not get it to work with my Seagate Barracuda ES SATA 3.0Gb/s 750-GB Hard Drive (ST3750640NS). The 750G Barracuda has worked fine for a year in my Dell PowerEdge T105, and I put FreeBSD 8.0 on it using my Dell OptiPlex 8400, where it also works fine. However, when I put the 750G Barracuda in the net5501, it does not get past POST. Most of the time I get to POST 0123456789bcefgh sometimes to POST 0123456789bcefghip and sometimes it even starts to check memory, but does not get past the Mbyte Once it acutally got past POST and into the ComBIOS, but the drive still not visible. This, to me, suggests a power issue. When my Soekris supplied wall warts started to go, I'd get behavior like that on my 5501 with just a CF card. Assuming you're powering the drive through the Soekris and it draws more power than the other two drives, it could be you're over taxing the power supply. You could try an experiment and try to find a way to power the drive externally. Hi Jed, Thanks for the suggestion! I am an absolute newbee on embedded and power issues ))-: According to the specs the SSD will draw 3.5 -4.2 Watts on operation. The Barracuda 7200.7 160Gb (which works) draws 12 Watts average on operation according to spec. The Barracuda ES 750Gb (which does not work) draws 13 Watts average on operation according to spec. However, according to specs both the Barracudas draw up to 2.8 amps on 12V, i.e. a mighty 33.6 Watts on spinup. So I guess the problem could be that the voltage is dropping during spinup. Everything is powered through a 12V, 1.5A C8 power supply (supplied by Cortex); i.e. 18 Watts. I have searched the Soekris specs, but I cannot figure out how much power the net5501 itself needs, neither how much power it can supply to the SATA disk. I think this was discussed on the list a little while ago. Maybe it got captured on the wiki, but it should be in the list archives. And that may have been one of those times Soren made a definitive statement. Would it make sense to purchase another external power supply (e.g. 12V, 3A), or does the Soekris net5501 internally limit the possible power to the SATA disk, now matter what the external power supply can deliver? I'd think that would be the next thing to try, assuming someone else doesn't come up with a definitive alternative solution. Obviously there are practical limits to the current going through the net5501, e.g. I'd think 200A might tend to convert circuit traces into blown fuses. And the net5501 web page does mention a 20W limit on the external DC input to the 5501. (I don't think I want to experiment with powering the drive externally) You could use the PC you built it on as a power supply, being careful with grounding. Borrowing an external disk enclosure would be better if that's an option. HTH, -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Installing FreeBSD 8.0 on Soekris net5501
[reply on list this time] This will cover the serial console issues, including where to put the -h: http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/serialconsole-setup.html -Jed On Wed Feb 24 22:39 , Jed Clear jcl...@speakeasy.net sent: On Wed Feb 24 16:47 , Henrik /KaarPoSoft hen...@kaarposoft.dk sent: Henrik /KaarPoSoft wrote: Scott Newell wrote: At 03:24 PM 2/24/2010 , Henrik /KaarPoSoft wrote: Dear all, I have just purchased a Soekris net5501 and tried to install FreeBSD 8.0 on it. However, I cannot get it to work, and I am hoping for your kind suggestions. Try booting with ACPI disabled? It was one of the many things that gave me fits when I tried to get 8.0 running on a '4521. I am not quite sure how to do that. Yes, in a normal FreeBSD boot I get the option, but here I just get a garbled terminal screen... OK, the garbled output sounds like one or several issues. As a previous poster said, the path of least resistance is to change the Soekris default BIOS 19200 console speed to 9600 (and your terminal emulator) as FreeBSD uses that speed in at least two different stages of the boot. Second is that you need to have a -h in some boot options file to use a serial console. I forget the exact file, but google should turn it up for you. Something in /boot/. Another issue is that a normal FreeBSD boot puts up a semi-graphic menu during the boot to choose among various options. I believe when you build the NanoBSD variant that the no beastie customization eliminates this. Lastly, you could have a problem with /etc/fstab. The device names in there probably reflect the device naming on your SERVER and must be changed to reflect the device naming on the Soekris box or your boot will fail when it attempts to mount the filesystems. This could result in no changes to any files on the disk as you're seeing. Have you considered using a CF card instead of a disk? The NanoBSD build of FreeBSD works rather well and you can add as much of the full FreeBSD functionality as you want and will fit on you CF card. If you don't have a lot of dynamic data, this is a better way to go. This is how I use my 5501. No moving parts and less heat. Or you could use NanoBSD and the CF card for the OS and Apps, and the HDD for just your dynamic data. HTH, -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Debian install on 4801, eth0 won't work.
On Tue Feb 23 12:59 , der Mouse sent: Not real sure that X would do a lot of good on a 4801 box I don't see why not. It was useful on Sun-3s, which were significantly wimpier. Perhaps you're confusing X with insanely bloated modern desktop environments built atop X, like KDE and Gnome? (X is not a window system; it's a framework for building window systems. It's possible to build fairly lightweight window systems in X.) I don't think the X _server_ software would be of much use on a Seokris box, not because of horsepower, but because of lack of KVM. No reason you couldn't run X client software (e.g. xterm) though. Although it seems like HTTP[S] based interfaces are much more common these days. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Writing a BIOS config from cmd line
Have you considered simply set bootdrive=F0 80 and then just control the boot from the boot server? You'd just need some way for the boot server to selectively ignore boot requests from the soekris box when you want the soekris to boot from its own drive. But since that server is presumably running an OS all the time, it seems like a much simpler problem. If a given boot server only supports a single soekris, this could be as simple as disabling the boot service entirely. If multiple soekris are supported you might do something with ACLs (e.g. firewall rules). On an old Sun server, I'd be editing /etc/ethers to do this. Another thought was that in ye olden days (i.e. before the Internet was so ubiquitous) we'd be sticking a modem on the console port. Obviously modems and phone lines aren't free, but it does give you complete remote console access. Note that these ideas were inspired by a few pints of Square One's Hop Fest, with the added distraction of the Phillies working hard to clinch the NL tonight, so YMMV. -Jed On Tue Oct 20 22:23 , Erik M. Cummings sent: I agree with the logic structures being a pain. This is how the current lab implementation is built (bash + expect) as a proof of concept, however, I cannot ADD another system to each one of these in the field because I need to change the boot order sometimes. I will need to get the package working from a linux shell-- or use alternative system boards. Right now the process of using flashrom to save and restore two different snapshots of the BIOS seems most promising. Erik On Oct 20, 2009, at 7:03 PM, The Fungi wrote: On Tue, Oct 20, 2009 at 08:45:57PM -0500, chad brabec wrote: [...] the lanuage is based off of tcl and you can get it with tk also. [...] Though logic and syntax in TCL are a pain, so I personally recommend a pure Python module which reimplments the interaction pieces allowing you to do the rest in a (granted, my opinion) much more pleasant language: http://pexpect.sourceforge.net/pexpect.html I've used it recently on a number of projects where I previously would have employed traditional expect instead, and been quite happy with the results. -- { IRL(Jeremy_Stanley); PGP(9E8DFF2E4F5995F8FEADDC5829ABF7441FB84657); SMTP(fu...@yuggoth.org); IRC(fu...@irc.yuggoth.org#ccl); ICQ(114362511); AIM(dreadazathoth); YAHOO(crawlingchaoslabs); FINGER(fu...@yuggoth.org); MUD(fu...@katarsis.mudpy.org:6669); WWW(http://fungi.yuggoth.org/); } ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Installing NetBSD-5.0.1 on net5501
On Tue Sep 22 17:50 , Manh Do sent: Hello I'm trying to install the official NetBSD-5.0.1 by using the official GENERIC kernel config file but then I ran into the following error message at bootup time:    warning: no /dev/console   panic: [snip] Any ideas on how to fix this problem? At a wild guess (from a FreeBSD user), you need to figure out how tell NetBSD to use the serial port as a console since most x86 OS assume a KVM style console. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Fail to boot with equipment connected to serial port
On Mon Jun 1 18:51 , Asle Festø sent: Hello, I'm a newbie trying to learn about these fascinating little devices. Sorry if my question is to easy but I haven't found the answer just yet. I have a Soekris Net4801 that I want to use for data logging. I have installed Debian on it and want to use it to read data from a 1-Wire network. This is done by the serial port. I do however have trouble booting with the 1-wire device attached to my box. It boots just fine without anything connected at all but if I connect the cable from my 1-wire device it fails to boot. I don't know if it's the BIOS or Debian/Grub that needs to be tweaked. You might find it easier to connect your one wire network to the second serial port. This avoids all the potential console difficulties, which would likely reoccur every time you upgrade the OS. This option also leaves your console port free for the inevitable debugging. The second serial port is on a 10 pin header on the board. Check the manual and wiki for more information. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Net5501 and Crux
On Sun Apr 12 21:57 , 'TechKid PC' tho...@techkidpc.com sent: From: Christopher R. Hertel c...@ubiqx.mn.org That's why it's called comBIOS. The Soekris BIOS sends its output to the serial port. Ah thanks that explains it. That was the only way I assumed would prove useful for trouble shooting any boot errors but all the documentation I have found online did not mention it operated in that manner. Last question I hope: Do you have any links that explains how to initiate a connection to the combios from another computer? I assume I use telnet or hyper terminal? What options should be used? http://wiki.soekris.info/Connecting_to_the_serial_console HTH, -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] impossible to link eth1, eth2,
On Fri Dec 5 20:25 , David Alexander sent: Several people here assume you want a bridge to tie all your interfaces together. You can go to the trouble of doing that, but why? The NET5501 is the first Soekris box that can actually saturate the 100Mbit link. With that in mind I had high hopes that bonding several nics together might make up for the fact that I was moving my server from a Gbit capable box to the Soekris. Alas, bonding the nics didn't provide any real throughput improvements, at least not according to my primitive tests. The 5501 can really only saturate a single link at a time. If someone else thinks otherwise I'm all ears. First do you understand the difference between bridging and aggregation? If you were trying to bridge multiple interfaces together to the same switch, or interconnected switches, spanning tree should put all but one into a blocked state, so you get no bandwidth increase, just a slow fail over path. Now if you set up a link aggregation (sometimes called by the ambiguous term trunking), you can increase your bandwidth in general use. However you have to be aware of how the hashing algorithm is set up that picks which link to use for any given packet. Often it hashes on the IP address pair. This is done intentionally to put any given flow on the same link to lower the probability of out of order packet delivery. But it also means that if you only test an aggregate with one pair of boxes, you're again only using one link. Some switches and devices support round robin, which should boost your bandwidth, but you risk out of order packet arrival. And after all that, I have no idea if the 5501 could saturate more than one link, just that under those circumstances it won't even try. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] FreeBSD on Geode
On Wed Nov 19 16:47 , PatrickLamaizière sent: Le Wed, 19 Nov 2008 21:26:43 +0100, Matthias Luft [EMAIL PROTECTED] a écrit : And this is exactly what I'm planning to do: building a small homeserver on a Geode LX. So my main concern is about the performance: Could anyone tell me the, how much throughtput on a 100Mbit network via scp is possible? This would be a great clue for me! With FreeBSD Current on my net5501: Around 3.0 MB/s, with a patch upon openssl to make it uses the AES engine of the geode : 5.5 MB/s I can confirm 3.2 MByte/sec scping a 3.3MB kernel file of a CF card. No patches, just a NanoBSD build with a custom minimized kernel. If I read 100MB out of /dev/zero with ssh, it goes down to 2.9. -Jed net5501# uname -a FreeBSD net5501 6.3-RELEASE-p4 FreeBSD 6.3-RELEASE-p4 #0: Mon Sep 29 21:05:32 EDT 2008 [EMAIL PROTECTED]:/usr/obj/nanobsd.custom/usr/src/sys/FOO i386 macmini2:~ foo$ scp [EMAIL PROTECTED]:/boot/kernel/kernel /dev/null Password: kernel100% 3277KB 3.2MB/s 00:01 macmini2:~ foo$ ssh -l foo net5501 dd if=/dev/zero count=102400 bs=1024 /dev/null Password: 102400+0 records in 102400+0 records out 104857600 bytes transferred in 35.934339 secs (2918033 bytes/sec) macmini2:~ foo$ ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] +5V power on Net4501
On Mon, 28 Jul 2008, Marc Balmer wrote: * Chris Babcock wrote: Can those LVC units be used for as a GPS reference time source for NTP? Yes. They output NMEA sentences, are fully (and easily) configurable and provide a PPS signal (which has to be turned on by sending the unit some commands). I have one doing that very job with a 5501. I use the $GPRMC sentence and turned of the rest. ISTR that was the sentnece recommended by the ntp.org s/w. If you'd rather not mess around with getting power from pin 10, and aren't otherwise using your USB connector, it is a +5V source, with enough hampsters for the LVC, at least with no internal HDD. Cutting up a spare USB cable is probably the easiest way to get connected there. This also lets you use a standard 2x5 to DE-9 cable for COM2:. Of course if you're putting a 2x5 header on the LVC cable anyway, never mind. If you haven't purchased the GPS 18 LVC yet, I seem to recall you could get it with two different length cables. I have 5m and I think the other option was 3m. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] +5V power on Net4501
On Mon, 28 Jul 2008, Jed Clear wrote: If you'd rather not mess around with getting power from pin 10, and aren't otherwise using your USB connector, it is a +5V source, with enough hampsters for the LVC, at least with no internal HDD. Cutting up a spare USB cable is probably the easiest way to get connected there. This also lets you use a standard 2x5 to DE-9 cable for COM2:. Of course if you're putting a 2x5 header on the LVC cable anyway, never mind. Forgot to add that I put DE-9 and USB connectors on my GPS 18 so I could move it between different PCs. Moving it around is necessary since Garmin updates firmware via MSWindows and my Soekris is MS free. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Squashed versions of FreeBSD
On Wed, 2 Jul 2008, Brett Glass wrote: Several years ago, I developed several embedded systems using PicoBSD, a squashed version of FreeBSD. I then was diverted from embedded UNIX development for a few years and am finally returning to it. I need to upgrade/revamp some of my older work and then develop some new things. I'd like to use a derivative of FreeBSD 7.0 (or the upcoming 7.1), and need to come up to speed on the various existing schemes for reducing FreeBSD 7.x to manageable size and putting it on a CF card. All of the systems I'll be developing will need to be configured remotely, and I may need to do remote firmware updates, so I'm especially interested in schemes which accommodate these things. I expect to build the systems around Soekris boards, PC-compatible set top boxes, or similar hardware. So far, I've seen two possibilities: miniBSD and a nanoBSD. Are there others of which I should be aware? What are their pros and cons? How does each of them handle issues involving nonvolatile/read-only media vs. read/write media? Are good howtos and architecture papers available for them, so that I do not have to reverse engineer them to fully understand how they work? Any input would be MUCH appreciated. I'm using nanoBSD on a 5501-70 and it works fine. I am using FreeBSD 6.3 as the base, though. And I didn't find any problems finding pointers on how to get it running. Probably the biggest problem I see with is that you have to roll your own to add non-base ports or packages. Not that it's hard to do, but it seems like a prime candidate to add to the stock nanoBSD build process. I did have the luxury of writiing the CF card on my build server the first time, but you should be able to leverage the PXEboot if you need to load the CF card with the 5501 the first time. nanoBSD is designed to only use half your CF at a time, allowing subsequent updated images to be loaded in the other half and then boot over when complete. This will get you rolling: http://www.freebsd.org/doc/en_US.ISO8859-1/articles/nanobsd/howto.html -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] asking about voltage and temperature monitoring soekris net 4801
On Fri, 23 May 2008, Wilkinson, Alex wrote: 0n Thu, May 22, 2008 at 06:41:06PM +0800, komang adi wrote: yes, i have seen that before. But, i'm using ubuntu for the OS. The program at http://phk/soekris/env4801/ using free besd. Can i just apply it for the ubuntu? Thanks Poul, the aforementioned URL redirects to http://www.searchfusion.com/ I suspect this is what he war referring to: http://phk.freebsd.dk/soekris/env4801/ -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Want to build a router
On Thu, 22 May 2008, jmc wrote: --- Stephen Brown [Thu, May 22, 2008 at 07:46:24PM -0400]: --- Well I'm looking to get away from my old PC and jump into an embedded platform because I think it will be challenging, fun and educational while fullfilling a need at the same time, and I already have a dedicated switch so the 4 port card wouldn't help me. amen, brother. all about the hack. For me it turned out to be all about the quiet. I got the 5501 to replace an ailing 486/66 tower as my home firewall. I planning to repair the 486 and use it for something else, even purchased the parts, but it is so much quieter in here with it off that I may never get to it. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] openbsd install problem
On Mon, 28 Apr 2008, Kjetil wrote: I am trying to install openbsd but I keep getting an error message, booting tftp:/bsd: open tftp:/bsd: Unknown error: code 60 failed(60). will try /obsd [snip] Try add these two options to your dhcpd.conf (i assume your tftp server is the same as your dhcp server). next-server 10.0.0.100; option root-path /bsd; Some [not so] obvious things to check: Is /bsd and all relevant contents world readable? Did you restart inetd after uncommenting tftpd in /etc/inetd.conf? Did you set up the tcpd wrapper properly? From the linux box can you do manual tftp? tftp localhost get bsd.rd -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net4801 + openbsd + serial device
On Sun, 27 Apr 2008, Uffe Jakobsen wrote: The part of the graphing script communicating with the device is this: printf \x11\x00\x00\x00\xB6\x00\x00\x00\xC7 /dev/tty dd if=/dev/tty of=/root/tmp/output bs=1 count=31 The result I get: Looking at the script part that you've provided in your email and looking at the supplied output I would say that you use the wrong tty device. The output that was supposed to go to the solar-device is output to your own terminal. Now I'm not that familiar with OpenBSD but /dev/tty is usualy the system concole device on most unix'es On Solaris/BSD's serial-devices/-ports are usually called something like /dev/cua?? /dev/ttyd? But maybe someone on this list that know OpenBSD better that I can help you with that actual device naming for serial devices on OpenBSD ? In FreeBSD, /dev/tty is what ever tty you're presently logged in on, won't even be a serial port at all if you're SSH'd in. You need something like /dev/ttyd0 (FreeBSD) to specify the first serial port. You probably also need to get OpenBSD to take the getty off that serial port. Not sure how far OpenBSD has diverged from FreeBSD, but in FreeBSD you'd edit /etc/ttys and find the line with ttyd0 and change the keyword on to off, save, then kill -1 1. One thing concerns me is that I've never run a Unix with no console device enabled, even if it was just an unconnected serial port. Another alternative is that there is a second serial port on the 4801. You just need to bring it out to a DB9. Beware there are two different 2x5 to DB9 wirings, beyond the usual DCE/DTE problems, and the nomenclature is horribly inconsistent. I went through all that to hook up my GPS18, getting the wrong cable the first try. I updated the Soekris wiki on this, so check there to avoid my mistakes. ;-) -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] SATA cables on a net5501: where to hide them
On Wed, 23 Apr 2008, jmc wrote: --- Wim Vandeputte - Mailinglists only [Wed, Apr 23, 2008 at 08:32:40PM +0200]: --- On Wed, Apr 23, 2008 at 12:18:35PM -0400, jmc wrote: my only remaining concern is that i'm afraid i have too tight a bend radius to keep the cable away from the PATA connectors which might rough up or even puncture the cable itself. but then again, this isn't going in a tank or off-road vehicle---i figure it should be OK, but i try to always be cognizant of bend radiuses with cables that carry data. If you're worried about the PATA pins I have a suggestion. Cut or remove the ribbon cable from the connector of a PATA cable and install just a connector on the pins. If you've been building PCs for a while, you probably have a few laying about in junk drawers. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net4826 lockup issue and BIOS release
On Fri, 18 Apr 2008, Izumi Utakoji wrote: Looking into ComBIOS changelog, we are considering to upgrade the BIOS, but want to make sure we can apply b4801_133.bin to net4826. I assume that net4801 bios can be applied to net4826, since it sais for net48xx, but want confirmation. Would also be useful if they updated the download page section title to ...for net48xx series. Also, since I am a new comer here, allow me for the basic question, but how do you do search through the archive at Soekris site??? Use Google with a search string like: 4826 update bios site:lists.soekris.com You can also browse the archive here: http://lists.soekris.com/pipermail/soekris-tech/ -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] CF often not detected by 5501-70 ComBIOS 1.33
On Sun, 13 Apr 2008, John Stalker wrote: I am experiencing problems with intermittent detection of my CF card on a new 5501-70 with the latest version of the ComBIOS. The card is a new Silicon Systems SiliconDrive Compact Flash card 2 GB Industrial Grade card, so it's not an old or a junky card. No sol'n, but some ideas for further data points: Both being new could be a bad thing. If you have any veteran hardware available, try testing with that. For instance, does another 2GB CF card boot reliably in the new 5501? Does the CF card in question work well with another reader, preferably another Soekris? When you PXE boot, is the CF card visible to the OS? If so, try to fsck it, or run an md5 on the device, to give it a good thrashing. Lather, rinse, repeat. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] ethernet problem
On Wed, 2 Apr 2008, [UTF-8] Lars Noodén wrote: Michael Greenish wrote: Any one seen this before with their Soerkis box? Any suggestions? Not that specifically. I have seen with a 4801 that some subnets I have to work with will not let me use a normal cross-over Ethernet cable with the Soekris unit. But those particular wall plugs do work with a straight-through (usually they're orange) cable as well as with a cross-over if I connect first to a hub and then to the 4801. That was my first thought also on reading the original post. The 4501 appears to be MDI only and likely the Linux box is as well while the router is either auto MDI or MDI-X. That would be consistent with the described symptoms, in that both work with the router and not with each other. Try an Ethernet crossover cable between the 4501 and Linux box or a hub and straight cables. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] FreeBSD 7.0 on a net4801?
On Sat, 1 Mar 2008, Jan Hoevers wrote: Graham Menhennitt wrote on 1-3-2008 7:56: Paul Hoffman wrote: Greetings again. I'm trying to install FreeBSD 7.0 on a net4801 with no luck. After the (unreadable over serial) boot prompt, during the 10-second countdown, I hit return. The cursor goes almost to the left end of the screen and stops. It doesn't respond to typing. Eventually, the system reboots. I had no problem installing a pre-release version of 7.0 a few months ago. Why is the serial output unreadable? The Soekris BIOS talks at 9600 baud and FreeBSD defaults to 19200 so you might need to change baud rates in mid boot. I also needed to disable DMA as a few recent messages have mentioned. But it doesn't sound like you're even getting that far. What are you installing from and to? It's the other way round actually. Soekris bios defaults to 19200 baud and FreeBSD to 9600. An easy solution is to set the bios to 9600, no need for acrobatics like changing baud rate in mid boot. Along the same lines, you need to tell FreeBSD to use a serial console. Otherwise it assumes you have a graphics adapter and keyboard attached. I forget off the top of my head the incantation needed, but it's in the Handbook. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] FreeBSD 7.0 on a net4801?
On Sat, 1 Mar 2008, Paul Hoffman wrote: At 5:56 PM +1100 3/1/08, Graham Menhennitt wrote: Every character typed by the FreeBSD boot process is erased after typing, at least using the console programs I have tried. It has always been this way. (I don't build my own kernels.) The Soekris BIOS talks at 9600 baud and FreeBSD defaults to 19200 so you might need to change baud rates in mid boot. No, it's not that. I see each character come up and disappear. As I said, I see the countdown happening. If your net4801 has the comBIOS serial speed set to 19200 (the default), at a certain point in the FreeBSD boot process, the default FreeBSD configuration will change to 9600. If your terminal [emulator] is still set to 19200, you may see garbage, but you may also see nothing. But it doesn't sound like you're even getting that far. What are you installing from and to? 7.0 RELEASE, from CD onto a CF card on a different machine. This might be a lot clearer if there was a simple way to get the FreeBSD boot stuff to leave its text on the serial console in a way that I could see. Am I missing some bit of passed-down wisdom here? It does, if you configure it properly to use a serial console, but that is not the default. The handbook section on Serial Console is here: http://www.freebsd.org/doc/en/books/handbook/serialconsole-setup.html When you do 24.6.2 step 3, also check the speed setting in /etc/ttys matches what you have in comBIOS and your terminal. Note you can do those commands while the CF card is mounted on your different machine, just prepend the mount point to the file paths mentioned. The handbood doesn't mention it, but you probably want to set dcons and all the ttyv0-8 to off in /etc/ttys. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Console connection
On Tue, 12 Feb 2008, voip crazy wrote: I planned to install Debian on the 5501 box but I'm not sure if install it on a CF card or in a SATA 2,5 Hdd drive. Which kind of install will be easier? which one would have less error or patching? How about the life of a CF card compare with the life of a SATA drive? Disk drive will likely be an easier install as typically you need a few extra steps with the CF card so that it is used read mostly. But there are enough examples out there that this shouldn't be a factor. All other things equal, the CF card should be more reliable as there are no moving parts and less heat. If you set up your CF read mostly, it should last indefinitely. If you need to write frequently to the CF (e.g. persistant logs), you will need to do more research. For starters, search this list's archives for the discussion on Flash write cycles. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
[Soekris] Nanobsd/FreeBSD 6.3-R won't boot
After updating my build machine to FreeBSD 6.3R, I created a new NanoBSD image for my net5501. I used the same nano and kernel config files as with 6.2R. 6.3R on the net5501 consistently fails the boot process while trying to mount / off of the CF card (see console below). After that I can't even get it to reboot off the old 6.2R partition, unless I pull the power. First off, has anyone else gotten 6.3R to boot on a net5501? I'm planning on trying a GENERIC kernel next, but am wondering if 6.3R is doing something different with DMA by default, perhaps needing the 1.32h and i comBIOS fixes. I think I have 1.32(release), at least it doesn't display any suffix letter when it boots: comBIOS ver. 1.32 20070606 Copyright (C) 2000-2007 Soekris Engineering. Any other suggestions? Thanks, -Jed Console output follows: --- Timecounter TSC frequency 499904773 Hz quality 800 Timecounters tick every 1.000 msec IPv6 packet filtering initialized, logging limited to 1 packets/entry ipfw2 (+ipv6) initialized, divert enabled, rule-based forwarding disabled, default to deny, logging limited to 5 packets/entry by default ad0: 977MB SanDisk SDCFB-1024 HDX 4.03 at ata0-master WDMA2 ad0: TIMEOUT - READ_DMA retrying (1 retry left) LBA=0 ad0: TIMEOUT - READ_DMA retrying (0 retries left) LBA=0 ad0: FAILURE - READ_DMA timed out LBA=0 ad0: TIMEOUT - READ_DMA retrying (1 retry left) LBA=0 ad0: TIMEOUT - READ_DMA retrying (0 retries left) LBA=0 ad0: FAILURE - READ_DMA timed out stray irq7 LBA=0 ad0: TIMEOUT - READ_DMA retrying (1 retry left) LBA=1 ad0: TIMEOUT - READ_DMA retrying (0 retries left) LBA=1 ad0: FAILURE - READ_DMA timed out LBA=1 ad0: TIMEOUT - READ_DMA retrying (1 restray irq7 try left) LBA=0 ad0: TIMEOUT - READ_DMA retrying (0 retries left) LBA=0 ad0: FAILURE - READ_DMA timed out LBA=0 Trying to mount root from ufs:/dev/ad0s1a Manual root filesystem specification: fstype:device Mount device using filesystem fstype eg. ufs:da0s1a ? List valid disk boot devices empty line Abort manual input mountroot ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] serial problem on new 5501
On Wed, 16 Jan 2008, Wim Kerkhoff wrote: Jed Clear said the following, On 1/15/2008 3:18 PM: On Tue, 15 Jan 2008, Wim Kerkhoff wrote: I have a new 5501 here. Serial has never been used. I've tried all the common speeds, 8N1 no flow control, but no luck. The cable is definitely fine since it works with a switch. The switch could be DCE pin out and 5501 DTE (or vice versa). Got a null modem to try in there? I'm in now with a real null modem cable, I should have known better then that. I have gotten many new pieces of kit this year at work, about 40% are DCE pin placement. I had to start taking notes on which RJ-45 adapter to use with which. I say placement because all of them use male pins, where only a DTE should be male. About the only piece of gear I know that gets it right is an external modem. It's invariably DCE and female. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] serial problem on new 5501
On Tue, 15 Jan 2008, Wim Kerkhoff wrote: Hello all I have a new 5501 here. Serial has never been used. I've tried all the common speeds, 8N1 no flow control, but no luck. The cable is definitely fine since it works with a switch. The switch could be DCE pin out and 5501 DTE (or vice versa). Got a null modem to try in there? Default speed is 19200. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Installing Debian on a 4501 with HDD on PCI Adapter
On Tue, 4 Dec 2007, Heather Lockridge wrote: From dmesg, I see: hda: max request size: 128KiB hda: 2001888 sectors (1024 MB) w/1KiB Cache, CHS=1986/16/63 hda: hda1 Which I presume is the on-board IDE controller. [snip] Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2 ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx IT8212: IDE controller at PCI slot :00:11.0 IT8212: chipset revision 19 it821x: controller in smart mode. IT8212: 100% native mode on irq 10 ide0: BM-DMA at 0xe020-0xe027, BIOS settings: hda:pio, hdb:pio ide1: BM-DMA at 0xe028-0xe02f, BIOS settings: hdc:pio, hdd:pio natsemi dp8381x driver, version 2.0, June 27, 2006 originally by Donald Becker [EMAIL PROTECTED] http://www.scyld.com/network/natsemi.html 2.4.x kernel port by Jeff Garzik, Tjeerd Mulder natsemi eth0: NatSemi DP8381[56] at 0xa000 (:00:12.0), 00:00:24:c8:cd:5c, IRQ 11, port TP. natsemi eth1: NatSemi DP8381[56] at 0xa0001000 (:00:13.0), 00:00:24:c8:cd:5d, IRQ 5, port TP. natsemi eth2: NatSemi DP8381[56] at 0xa0002000 (:00:14.0), 00:00:24:c8:cd:5e, IRQ 9, port TP. Done. Begin: Mounting root file system... ... Begin: Running /scripts/local-top ... Done. Begin: Waiting for root file system... ... root file system never mounts. Caveat, I am not a Linux person, but have watched SunOS and FreeBSD booting way too many times. My impression on reading this is that the entries in /etc/fstab don't match the devices assigned between the BIOS and kernel. It looks to me like you are getting through the boot loader and the kernel has enumerated the devices, so you must be reading a drive (assuming I didn't lose track and you were net booting). And it just seems like the right time in the boot process that /etc/fstab has just come into play. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] 1GB Compact Flash 6.06 Dapper Soekris
On Thu, 29 Nov 2007, ervin wrote: I'm pxe booting ubuntu dapper 6.06 to a Soekris 4501 (64MB RAM 1GB Compact Flash card) ... dhcp, tftp and the server install is running smoothly until I want to write the partitioning to the CF card I tried guided and manual partitioning same issue, which is: issue - The attempt to mount a file system with type swap in IDE1 master, | | partition #5 (hda5) at none failed. -- I made tests on different CF card 1GB 4GB and from both Sandisk IBM ... no difference. My manual partitions: 200 MB / 100 MB swap 350 MB /usr 350 MB /var I'm a BSD person, so I won't speculate on exactly what's going on, but it seems like this is protecting you from your own folly. Swap is the last thing you want on a flash disk. Flash has a limited number of writes before it degrades. ISTR that the spec for write cycles in in the 100s of thousands, which is fine for a digital camera, but a swap partition will chew through that in no time. Suggest you find a HOWTO specifically for a flash filesystem, which you will probably find has no swap at all. If you really need to have virtual memory for your application, you could consider a CF format microdrive, or an IDE laptop drive. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
[Soekris] net5501 COM2 +5V pin
Can someone confirm that the net5501 has the same COM2 pin out as the net4801? If so, can that +5V pin supply at least 60 mA? I've got the 12V 1.5A wall wart supply and no rotating storage or PCI cards, just the CF card. If you're curious I'm considering connecting a Garmin GPS 18 LVC-5m. Tips on getting this working with ntpd in FreeBSD 6.2 are also welcome, if perhaps slightly off topic. Although I expect my biggest challenge will be getting enough of a view of the sky with all the aluminum siding Thanks, -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] NanoBSD question...
On Mon, 5 Nov 2007, Bill Borton wrote: I'm trying to install various programs from Ports into my NanoBSD Build. I'm looking at the following as a starting point for doing much the same (e.g. NTP from ports), but haven't started in on that project yet. http://www.nabble.com/Building-NanoBSD-with-ports-and-packages-t3614023.html I say starting point because I think it could be a bit cleaner. I recall finding some solutions that were better integrated but got distracted and didn't save any links beyond that one. Anyway I think they all had the same characteristics: build recursive packages from ports, null mount the package directory within the NANO root, then chroot and install packages. Side thought, with NTP you should remember to save off your drift file periodically. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Net5501 doesn't boot from HDU
On Tue, 30 Oct 2007, Marco A. Calamari wrote: I configured Linux+Grub on the 5501 booting without hassle from a 1GB CF. Today I added a tested Hitachi HDU with the ATA mounting kit The bios is a 1.32i and the HDU is cable select, as describet in a thread of this list. Pri Mas Flash Card LBA Xlt 1007-32-63 1015 Mbyte Pri Sla Hitachi HTS541680J9AT00 LBA Xlt 1024-255-63 78 Gbyte I seem to recall reading of some limitation where you could use CF or ATA, but not both. Not certain it was the net5501 either. But check the archives for Soren's posts. Or just try removing the CF. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Max MTU size
On Mon, 29 Oct 2007, Jeff Quast wrote: I thought jumbo frames (gigabit nics) were for lessening interrupts. Yes, and less checksum calculations. However interrupt coalescing and TCP Offload Engines make this less of an issue than it once was. A large frame that unpacks to many smaller frames by software (kernel). No, typically any size frames are assembled back into one large logical byte stream. It is not really 1500 bytes on the wire, is it? Not a simple answer. If your Path MTU supports 9k your packets remain 9k. More likely you will encouter a 1500 byte MTU subnet and either have your 9k packet fragmented or have to go throught the PMTUD process and end up with 1500 byte packets from end to end. But if you were asking if the kernel or NIC breaks them down from 9k to 1500 all the time, then no. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Max MTU size
On Sun, 28 Oct 2007, Jan Hoevers wrote: Thomas Elsgaard wrote on 28-10-2007 8:19: Has anyone tried to use a MTU size above 1500 bytes on the soekris 5501 ? I am looking for a device that can run with a MTU size of 1544 Maximum ethernet frame size is 1548 if I recall well. An IP packet with a size of 1500 meets that maximum. The Via 6105M Chips on the 5501 support 802.1q VLAN tagging, which would be an Ethernet frame size of 1522, maintaining the IP packet size of 1500. Since there is no mention of jumbo frames, that is likely it. However the link to the data sheet is 404'd, so we can't rule out some wiggle room. Not sure why you're looking for 1544, but unless every device in the path supports it, it may be futile. Have you tried to get PMTUD working? -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Max MTU size
On Sun, 28 Oct 2007, der Mouse wrote: The Via 6105M Chips on the 5501 support 802.1q VLAN tagging, which would be an Ethernet frame size of 1522, As I understand it, the maximum standard Ethernet frame is 1516 bytes: 6 bytes of destination MAC, 6 of source MAC, 2 of frame type, 1500 of payload, and 2 of FCS. 802.1q increases this by 4 (2 for the vlan-tagged frame type and 2 for the vlan number, basically), bringing it to 1520, not 1522. Where's the mistake? 4 bytes for FCS. And in the interest of full disclosure, there are a preamble and SFD that add 8 more bytes berfore the DA. http://www.javvin.com/protocolEthernet.html -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] Project Feasibility
On Mon, 22 Oct 2007, Joel Jaeggli wrote: Carlos M. Perez wrote: 3 ? Remote Console ? need something that we can run GNOME/Desktop on so that we can remote in, and use the device?s local browser (Firefox/Opera) to configure network devices ranging from Copiers, printers, IP phones, access points. If it has a web based GUI we need to access it. That's actually something that's better done with ssh port forwarding or an out and out ip tunnel of some sort (ip-in-ip l2tp ipsec etc) leaves you with a much lighter weight device and the browser runs locally on your own machine. for a concrete recipe example try this: ssh -l username bastion-host-name -L 8000:remotedevice-ip:80 which will bind local port 8000 to port 80 in a device accessible from the other end of the tunnel to your bastion host. If you have more than one or two end destinations at a given site, you might want to extend the above solution a bit. Run some sort of small, non-caching, http proxy on the net and point the SSH tunnel at that. Then configure your local browser to use localhost:8000 as a proxy. You should then be able to surf to all the remote web configs at a site w/o rebuilding the tunnel for each. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] net4501 problem
On Sat, 20 Oct 2007, ®Âµ¦ ®ÂmOs wrote: I try to install monowall on net4501, but I have a problem the network ligth never on. I try with different cable but nothing. I don't think the 4501 has auto-MIDIX ports, so you must use a crossover cable if you are connecting to another non-switch device, such as some DSL bridges or Cable modems. If you connect your 4501 temporarily to a switch with that cable, and the LED comes on, that's likely the problem. Googling ethernet crossover cable would be your next move for either DIY or purchase. -Jed ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech