Re: [Soekris] 4501 + OpenBSD -STABLE: fast installs, slowwwwwww upgrades
Good question, but I think you'll have better chances to have this problem/question solved on the openbsd mailing list. http://lists.openbsd.org/ Otfried --OpenBSD4.0 on net5501-- Hey folks, I have a 4501 running the latest ComBIOS (1.31b), booting OpenBSD 4.1-STABLE on a CF card. Everything works great. However, something I'm very curious about... When I choose the (I)nstall option in the OpenBSD installer, I can install a new -STABLE build of the OS from the CF card in about 24 minutes. That time includes formatting the CF and everything. If I choose (U)pgrade, it takes over 3 *hours* to install the same -STABLE build off the same CF card. As the tarballs are installing, I get the usual progress meter, but the larger ones time out over and over again, saying -- stalled --. Keep in mind, each time I test this, I have copied the entire distribution to the CF card, and have selected disk as the source of the tarballs. Nothing's going over the network whatsoever. Again, everything works perfectly when the OS is running. But I'm really, really curious to know why this happens. It _always_ completes successfully, the OS always runs fine, and I see no instabilities. I'd appreciate it if someone could give it a think, and offer any reasons... Thanks much! Benny ps: Dmesg follows: [EMAIL PROTECTED] ~]$ dmesg OpenBSD 4.1-stable (GENERIC) #25: Sun Aug 19 06:05:51 CDT 2007 [EMAIL PROTECTED]:/usr/src/sys/arch/i386/compile/GENERIC cpu0: AMD Am486DX4 W/B or Am5x86 W/B 150 (AuthenticAMD 486-class) cpu0: FPU real mem = 66678784 (65116K) avail mem = 52539392 (51308K) using 844 buffers containing 3457024 bytes (3376K) of memory mainbus0 (root) bios0 at mainbus0: AT/286+ BIOS, date 20/70/14, BIOS32 rev. 0 @ 0xf7840 pcibios0 at bios0: rev 2.0 @ 0xf/0x1 pcibios0: pcibios_get_intr_routing - function not supported pcibios0: PCI IRQ Routing information unavailable. pcibios0: PCI bus #0 is the last bus bios0: ROM list: 0xc8000/0x9000 cpu0 at mainbus0 pci0 at mainbus0 bus 0: configuration mode 1 (bios) elansc0 at pci0 dev 0 function 0 AMD ElanSC520 PCI rev 0x00: product 0 stepping 0.1, CPU clock 133MHz, reset 0 gpio0 at elansc0: 32 pins sis0 at pci0 dev 18 function 0 NS DP83815 10/100 rev 0x00, DP83815D: irq 10, address 00:00:24:c0:04:ec nsphyter0 at sis0 phy 0: DP83815 10/100 PHY, rev. 1 sis1 at pci0 dev 19 function 0 NS DP83815 10/100 rev 0x00, DP83815D: irq 11, address 00:00:24:c0:04:ed nsphyter1 at sis1 phy 0: DP83815 10/100 PHY, rev. 1 sis2 at pci0 dev 20 function 0 NS DP83815 10/100 rev 0x00, DP83815D: irq 5, address 00:00:24:c0:04:ee nsphyter2 at sis2 phy 0: DP83815 10/100 PHY, rev. 1 isa0 at mainbus0 isadma0 at isa0 pckbc0 at isa0 port 0x60/5 pckbd0 at pckbc0 (kbd slot) pckbc0: using irq 1 for kbd slot wskbd0 at pckbd0: console keyboard wdc0 at isa0 port 0x1f0/8 irq 14 wd0 at wdc0 channel 0 drive 0: SanDisk SDCFH-2048 wd0: 4-sector PIO, LBA, 1953MB, 4001760 sectors wd0(wdc0:0:0): using BIOS timings pcppi0 at isa0 port 0x61 midi0 at pcppi0: PC speaker spkr0 at pcppi0 npx0 at isa0 port 0xf0/16: reported by CPUID; using exception 16 pccom0 at isa0 port 0x3f8/8 irq 4: ns16550a, 16 byte fifo pccom0: console pccom1 at isa0 port 0x2f8/8 irq 3: ns16550a, 16 byte fifo biomask f3c5 netmask ffe5 ttymask ffe7 pctr: no performance counters in CPU dkcsum: wd0 matches BIOS drive 0x80 root on wd0a rootdev=0x0 rrootdev=0x300 rawdev=0x302 ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] FreeBSD net4801
Tuomo Latto wrote: Does anybody happen to know the status of these http://www.soekris.com/Issue0003.htm with regard to Freebsd 6-STABLE? How much manual patching is required? Are the instructions in http://63.249.85.132/geode.html still needed? How valid are they? I suspect I might be hitting the 64K transfer issue (hard hang, no other hints whatsoever) with jailed, nullfs mounted bitflu when running (unmodified) 6.x from 2.5 drive (57231MB HTS548060M9AT00 MGBOA53A at ata0-master UDMA33). Forgot to mention that I'm running comBIOS 1.30. I also seem to be hitting the issue with a simple make buildworld. That should be a little more easier for other people to test. Although I admit, that might not be any easier to debug. As for what I've ruled out... Not much. It can't be the power, since I've also tried with another power source (12V/5A). I've run smartctl selftests and checked the SMART error logs. Nothing could be found there. Overheating? Possible, but the box isn't *that* hot to touch. It's merely warm. I don't know how useful this is, but after the hang if I don't do a cold boot (by holding the reset button pressed for awhile), the HD isn't shown on boot messages and the board tries a PXE boot. When the PXE boot doesn't work, the comBIOS console appears. After a cold boot (reboot) the HD is once again found. (If it wasn't for the Fujitsu drive in the original post on the 5501 reboot problem thread, I might see a pattern there.) Anything else I should know about? Or any more details I should be telling you? -- Tuomo ... And don't start a sentence with a conjunction ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] What to do with my old 4501
Hi, (1) (Almost) identical config for load-balancing and/or cases where the 4801 breaks down or needs to be pulled down for maintenance. CARP is tailored for such setups, but I couldn't get it working together with SSH (that was on OpenBSD 3.9, using the local ethernet interface for SSH logins - with a separate maintenance network it should give no problems). (2) In case you are offering external SSH access to any live device: set up the 4501 as an SSH RAS for your network. You'll have to log in twice or more to reach a particular machine, but at least it won't have to be directly exposed to the Big Bad World anymore. And you can play all kinds of tricks to make a cracker's life more difficult - fun!! (3) Use it for controlling a robot. Would require some real-time capabilities from the OS though. But if you've ever programmed an AVR directly, you'll know how incredibly comfortable high-level languages running with MB's rather than kB's of memory are. I'd very much like to give this a try, in case anyone's interested. Should provide Soekris Engineering with a great boost too. Robots are hot! (it seems). Bill On Sat, 2007-08-25 at 10:14 -0500, Ronald L. Rosson Jr. wrote: I currently have a NET4501 sitting on the shelf doing nothing since it was replaced by my NET4801 as a firewall running pfSense. I am looking for a low cost solution to have this system do something that can benefit my network that the NET4801 is not doing at the moment. any idea? TIA -Ron ___ 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] What to do with my old 4501
I would like to hear about some of these tricks. Its sounds like fun. I have also thought about controlling a robot with a soekris, but wouldent you need more GPIO? or would you use a simple AVR program to drive the servos etc? On 8/27/07, Bill Maas [EMAIL PROTECTED] wrote: Hi, (1) (Almost) identical config for load-balancing and/or cases where the 4801 breaks down or needs to be pulled down for maintenance. CARP is tailored for such setups, but I couldn't get it working together with SSH (that was on OpenBSD 3.9, using the local ethernet interface for SSH logins - with a separate maintenance network it should give no problems). (2) In case you are offering external SSH access to any live device: set up the 4501 as an SSH RAS for your network. You'll have to log in twice or more to reach a particular machine, but at least it won't have to be directly exposed to the Big Bad World anymore. And you can play all kinds of tricks to make a cracker's life more difficult - fun!! (3) Use it for controlling a robot. Would require some real-time capabilities from the OS though. But if you've ever programmed an AVR directly, you'll know how incredibly comfortable high-level languages running with MB's rather than kB's of memory are. I'd very much like to give this a try, in case anyone's interested. Should provide Soekris Engineering with a great boost too. Robots are hot! (it seems). Bill On Sat, 2007-08-25 at 10:14 -0500, Ronald L. Rosson Jr. wrote: I currently have a NET4501 sitting on the shelf doing nothing since it was replaced by my NET4801 as a firewall running pfSense. I am looking for a low cost solution to have this system do something that can benefit my network that the NET4801 is not doing at the moment. any idea? TIA -Ron ___ 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 -- If you are savvy and smart about the choices you make in life, The sky is not the limit! Mark Shuttleworth dshufelt Well I'm not traveling with out Ubuntu CD's and my laptop. I may need to host a mini in flight install fest. 09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0 ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] 4501 + OpenBSD -STABLE: fast installs, slowwwwwww upgrades
That's because openbsd mounts the filesystem in asychronous mode on install (like linux ext2fs runs all the time) but does not on upgrade (so that you don't potentially lose significant amounts of data in a crash) If you aren't using asychronous mounts, a lot more data gets written when the filesystem changes. Combine this with the Elan SC420's PIO mode for data transfers and you experience 3 hours of suffering, like you describe. If you can create a complete disk image that you write to the flash on upgrade, you will have a better experience. If you must use your flash like a hard disk, try and upgrade to a 4801 or 5501 where you can use DMA modes with the CF (and find a suitable high-speed flash that implements it, which yours might alreay do.) By the way, older 4801s don't have DMA wired to the CF port. C. Bensend [EMAIL PROTECTED] wrote: Hey folks, I have a 4501 running the latest ComBIOS (1.31b), booting OpenBSD 4.1-STABLE on a CF card. Everything works great. However, something I'm very curious about... When I choose the (I)nstall option in the OpenBSD installer, I can install a new -STABLE build of the OS from the CF card in about 24 minutes. That time includes formatting the CF and everything. If I choose (U)pgrade, it takes over 3 *hours* to install the same -STABLE build off the same CF card. As the tarballs are installing, I get the usual progress meter, but the larger ones time out over and over again, saying -- stalled --. Keep in mind, each time I test this, I have copied the entire distribution to the CF card, and have selected disk as the source of the tarballs. Nothing's going over the network whatsoever. Again, everything works perfectly when the OS is running. But I'm really, really curious to know why this happens. It _always_ completes successfully, the OS always runs fine, and I see no instabilities. I'd appreciate it if someone could give it a think, and offer any reasons... Thanks much! Benny ps: Dmesg follows: [EMAIL PROTECTED] ~]$ dmesg OpenBSD 4.1-stable (GENERIC) #25: Sun Aug 19 06:05:51 CDT 2007 [EMAIL PROTECTED]:/usr/src/sys/arch/i386/compile/GENERIC cpu0: AMD Am486DX4 W/B or Am5x86 W/B 150 (AuthenticAMD 486-class) cpu0: FPU real mem = 66678784 (65116K) avail mem = 52539392 (51308K) using 844 buffers containing 3457024 bytes (3376K) of memory mainbus0 (root) bios0 at mainbus0: AT/286+ BIOS, date 20/70/14, BIOS32 rev. 0 @ 0xf7840 pcibios0 at bios0: rev 2.0 @ 0xf/0x1 pcibios0: pcibios_get_intr_routing - function not supported pcibios0: PCI IRQ Routing information unavailable. pcibios0: PCI bus #0 is the last bus bios0: ROM list: 0xc8000/0x9000 cpu0 at mainbus0 pci0 at mainbus0 bus 0: configuration mode 1 (bios) elansc0 at pci0 dev 0 function 0 AMD ElanSC520 PCI rev 0x00: product 0 stepping 0.1, CPU clock 133MHz, reset 0 gpio0 at elansc0: 32 pins sis0 at pci0 dev 18 function 0 NS DP83815 10/100 rev 0x00, DP83815D: irq 10, address 00:00:24:c0:04:ec nsphyter0 at sis0 phy 0: DP83815 10/100 PHY, rev. 1 sis1 at pci0 dev 19 function 0 NS DP83815 10/100 rev 0x00, DP83815D: irq 11, address 00:00:24:c0:04:ed nsphyter1 at sis1 phy 0: DP83815 10/100 PHY, rev. 1 sis2 at pci0 dev 20 function 0 NS DP83815 10/100 rev 0x00, DP83815D: irq 5, address 00:00:24:c0:04:ee nsphyter2 at sis2 phy 0: DP83815 10/100 PHY, rev. 1 isa0 at mainbus0 isadma0 at isa0 pckbc0 at isa0 port 0x60/5 pckbd0 at pckbc0 (kbd slot) pckbc0: using irq 1 for kbd slot wskbd0 at pckbd0: console keyboard wdc0 at isa0 port 0x1f0/8 irq 14 wd0 at wdc0 channel 0 drive 0: SanDisk SDCFH-2048 wd0: 4-sector PIO, LBA, 1953MB, 4001760 sectors wd0(wdc0:0:0): using BIOS timings pcppi0 at isa0 port 0x61 midi0 at pcppi0: PC speaker spkr0 at pcppi0 npx0 at isa0 port 0xf0/16: reported by CPUID; using exception 16 pccom0 at isa0 port 0x3f8/8 irq 4: ns16550a, 16 byte fifo pccom0: console pccom1 at isa0 port 0x2f8/8 irq 3: ns16550a, 16 byte fifo biomask f3c5 netmask ffe5 ttymask ffe7 pctr: no performance counters in CPU dkcsum: wd0 matches BIOS drive 0x80 root on wd0a rootdev=0x0 rrootdev=0x300 rawdev=0x302 ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
Re: [Soekris] 4501 + OpenBSD -STABLE: fast installs, slowwwwwww upgrades
That's because openbsd mounts the filesystem in asychronous mode on install (like linux ext2fs runs all the time) but does not on upgrade (so that you don't potentially lose significant amounts of data in a crash) If you aren't using asychronous mounts, a lot more data gets written when the filesystem changes. Combine this with the Elan SC420's PIO mode for data transfers and you experience 3 hours of suffering, like you describe. If you can create a complete disk image that you write to the flash on upgrade, you will have a better experience. If you must use your flash like a hard disk, try and upgrade to a 4801 or 5501 where you can use DMA modes with the CF (and find a suitable high-speed flash that implements it, which yours might alreay do.) By the way, older 4801s don't have DMA wired to the CF port. Ah, OK. Thanks much, Chris! That explains it. :) I can deal with it, it's not a huge thing. I just just very, very curious. Thanks for enlightening me. Benny -- It makes me want to crawl back up the line and hit them in the face with a spade.-- Loren Wilton ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
[Soekris] Problems (probably geometry) with 16GB flash and 5501
I have a sandisk elite 3 16GB flash card for my 5501. A 2GB card worked fine, so I'm confident the box is OK. When I boot the 16GB card, I get the geometry printed as LBA xlat 1024--63 16007MB. Which seems OK. But FreeBSD started on sector 16 for slice 1 (ad0s1a begins on 16), and I would've expected it to start on 63. THe problem is the boot loaded, I get the OK: prompt, but it says it can't load kernel. If I try to ls the /boot directory, it tells me that /boot is not a directory. Take the CF card out, stick it in my notebook runnign FreeBSD, and yes it is... Any suggestions? I can change it over to not use slices, and just use the whole CF card, no real reason for a slice on a dedicated card I guess... ___ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech