 |
|||
HomeGraffitiWebCamAboutSiteVisualDevGPG Key |
Welcome to Orb Graffiti, a place for me to write daily about life and computers. Contrary to popular belief, the two are not interchangeable. EMAIL - I publish email sometimes. If you send me an email and you want privacy or anonymity, please say so clearly at the beginning of your message.. |
MONDAY
Tues
Wed
Thu
Fri
Sat
Sun
October 14, 2002 - Updates at 0900 EDT
Good morning. It's Columbus Day. This is the first time I've worked for a company that takes Columbus Day as a holiday. So here I am, slept in again for a third consecutive day. How lovely. Since it's a new day, I'll be successful in my final push to the wireless summit, right? Right?
gryphon root #tail -f /var/log/everything/currentOct 14 09:32:01 [cardmgr] socket 0: Netgear MA401RA Wireless Adapter Oct 14 09:32:01 [cardmgr] executing: 'modprobe hermes' Oct 14 09:32:01 [cardmgr] executing: 'modprobe orinoco' Oct 14 09:32:01 [cardmgr] executing: 'modprobe orinoco_cs' Oct 14 09:32:02 [cardmgr] executing: './network start eth1' Oct 14 09:32:02 [cardmgr] + [32;01m* [0m Bringing eth1 up... Oct 14 09:32:02 [kernel] eth1: Error -110 setting multicast list. Oct 14 09:32:02 [kernel] eth1: Error -110 setting multicast list. Oct 14 09:32:02 [kernel] eth1: Error -110 setting multicast list. Oct 14 09:32:02 [cardmgr] + [A [-7G [34;01m[ [32;01mok [34;01m] [0m gryphon root #iwconfig eth1eth1 IEEE 802.11-DS ESSID:"Wireless" Nickname:"gryphon.orbdesigns.com" Mode:Managed Frequency:2.447GHz Access Point: 00:30:AB:1D:16:06 Bit Rate=2Mb/s Tx-Power=15 dBm Sensitivity:1/3 Retry min limit:8 RTS thr:off Fragment thr:off Encryption key:off Power Management:off Link Quality:68/92 Signal level:-30 dBm Noise level:-144 dBm Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0 Tx excessive retries:0 Invalid misc:0 Missed beacon:0 gryphon root #ifconfig eth1eth1 Link encap:Ethernet HWaddr 00:09:5B:24:0F:2F inet addr:192.168.1.17 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:100 RX bytes:0 (0.0 b) TX bytes:0 (0.0 b) Interrupt:3 Base address:0x180
I mean, what's wrong with that? Nothing at all. Except that it doesn't actually work, as such. The ESSID matches the WAP (also a Netgear, an MR314 model), as does the frequency (which corresponds to Channel 8). The bit that worries me is those oddball [cardmgr] output lines in the logfile. Those I don't understand. Also, I think I'd better look into interrupt conflicts. (3) strikes me as one that might be in use by another device. That can certainly put a hink in things. If anyone sees anything interesting in there that I should look at, let me know, please.
Have a great day, happy Monday to you, more later when (and if) I succeed.
1017 - Oh, and sorry. I did put up the post at the time noted. However, I forgot to update the current and sitemap pointers to the new week's page. OFIM.
Mon
TUESDAY
Wed
Thu
Fri
Sat
Sun
October 15, 2002 - Updates at 0658
Good morning. Greg asked some very pointed questions at some time during the afternoon yesterday. Among them, <gasp> did the hardware work under Windows. Since I now have a spare spindle for the laptop, thanks to Tom, I popped on my copy of Win2K, and it took me all of 2 minutes to get the wireless card going. Two minutes. Some things are still definitely easier under Windows One must also admit that manufacturer's support of an operating system makes a big difference. Netgear doesn't have the word "Linux" on their website, so far as I can tell.
Anyway, that done, I know now that the hardware works. Then I put back in the drive, built a fresh kernel and set of drivers, including a patch that I found for putting the card in monitor mode, so that tcpdump works and I could see what was going on. Then I invited Greg into the box (when he became available) to help me try some of the debug steps that he and Tom had recommended. While I waited for him to show up, it rolled around to dinner time. Yummy stir-fry made by Marcia.
After dinner and dishes, I popped back upstairs to find Greg in IRC and in Gryphon. We mucked about with looking at files and settings for routes and setup and such, unplugged the wire, and wooo Hooooo! Wireless was happening! One day soon I'll tear it all down, and make it work from scratch, just to confirm and document the steps that work for me. I also wonder if the run under Windows initialized something in the card, and thereafter it was ready to work. Mmmmm. Tom and Greg rock, however. Both helped immensely in my quest to get this stuff working, through both moral and technical support. And yes, it shouldn't be that hard. There probably was a better choice of card, neh? But this one is supported, according to the docs. And it is. Maybe having a wired interface in the system confused things and made my job harder? Still, it just worked under Windows, first try.
By the way, Sally's doing pretty well, for an unhappy little doggy. She's done with her pain meds, and starting to take food again. Yesterday was her first day of any reasonable food intake at all, and we're pretty grateful. It was starting to become worrisome. She's not getting aroud much yet, and when she does, it obviously hurts some. But that will ease over the coming days. Then it'll just be uncomfortable, and she'll start looking forward to the soft cast ... until she figures out that it'll take another round of general anesthesia to get there. Poor baby.
Off to work with me now. See you around.
Mon
Tues
WEDNESDAY
Thu
Fri
Sat
Sun
October 16, 2002 - Updates at 0845
Good morning. It's a very rainy day here in greater Washington Metro. And last night, when I got home, wireless LAN didn't work again. Then it did. And then it didn't after a reboot. I stopped in disgust last night. This morning, the card was up. WTF???
Other than that, not very much interesting is going on, so let's have some quick takes: Sally continues to improve in appetite and demeanor. They haven't yet caught the sniper. There's a new OpenSSH release, not driven (for once) by a vulnerability. Security and feature enhancements only, according to the changelog. The funny thing about that one is that the announcement's been on the OpenSSH page since at least early Monday morning. But the actual release of code was only there for RPMS for at least 24 hours. How... odd. I'll wait for a couple of weeks and see how things shakeout before upgrading, myself.
OK, nose meets grindstone. See y'all later.
Mon
Tues
Wed
THURSDAY
Fri
Sat
Sun
October 17, 2002 - Updates at 0730
Howdy. Let me state this for the record. I had nothing to do with it. Really I didn't. The card just started working. There are caveats, but wireless Linux is now alive here, and this is how it works... First, after many, many loops through all the same steps, what I've settled on are the vanilla kernel drivers. They didn't work the first five times that I tried that, but gift horses in the mouth I will not look. Here are the pertinent parts of the kernel setup:
General setup --->
PCMCIA/CardBus support --->
<M> PCMCIA/CardBus support
[*] CardBus support
[*] Databook TCIC host bridge support
[*] i82092 compatible bridge support
[*] i82365 compatible bridge support
Network device support --->
Wireless LAN (non-hamradio) --->
[*] Wireless LAN (non-hamradio)
. . .
<M> Hermes chipset 802.11b support (Orinoco/Prism2/Symbol)
. . .
<M> Hermes PCMCIA card support
Once configured, a quick compile suffices:
# make dep
# make clean bzImage modules modules_install
For the utilities, cardmgr and such, I've got pcmcia-cs-3.2.1 installed from sources. The source tree is unpacked in /usr/src, parallel to the Linux source tree. I just unpacked the tarball there, typed make config, accepted all the defaults, then make all ; make install. The make install put a stock /etc/init.d/pcmcia file in place. Also, I've got these items configured. In the pertinent excerpt from the file /etc/pcmcia/wireless.opts:
. . .
# Generic example (decribe all possible settings)
*,*,*,*)
INFO=""
# ESSID (extended network name) : My Network, any
ESSID="Wireless"
# NWID/Domain (cell identifier) : 89AB, 100, off
NWID=""
# Operation mode : Ad-Hoc, Managed, Master, Repeater, Secondary, auto
MODE="Managed"
. . .
There I only changed two items, adding the ESSID setting to correspond with what I've got configured in the Netgear MR314 WAP. And I set the mode to Managed, so that all other settings would get picked up properly from the WAP. Then I made the following changes in /etc/pcmcia/network.opts:
# Host's IP address, netmask, network address, broadcast address
IPADDR="192.168.1.17"
NETMASK="255.255.255.0"
NETWORK="192.168.1.0"
BROADCAST="192.168.1.255"
# Gateway address for static routing
GATEWAY="192.168.1.1"
There's also the option, instead, to use DHCP in that file, but I like statically configured networks where possible. That's it. Now the card works. I'll have you know that I did these steps in this rough order at least three times, interspersed among working with pcmcia-cs drivers, linux-wlan-ng drivers, new kernels, other distributions. I can't begin to explain why it didn't work the first time, and did, this time. I can still swamp the card with data and take it offline by doing something like a really large rsync operation (.25 gig or more, at max transfer rate). Then I just type (as root) /etc/init.d/pcmcia restart. But for normal usage, email, browsing and other stuff I do all the time around here, it seems to be reliable. And to put the cherry on top of this mini-howto for my future reference, here's the output of lsmod (and yes, my wired ethernet interface is also build as a module, but is set to autoload at boot).
gryphon pcmcia # lsmod
Module Size Used by Not tainted
orinoco_cs 4168 1
orinoco 27160 0 [orinoco_cs]
hermes 5540 0 [orinoco_cs orinoco]
ds 6088 1 [orinoco_cs]
yenta_socket 8736 1
pcmcia_core 35040 0 [orinoco_cs ds yenta_socket]
eepro100 18104 1
Today, I've got a meeting down in the District, to set up a machine and do a bit of network testing. But I've got a couple of hours left before I need to leave for that one, so I'll spend a moment and pick up some of the mess my ranting and raving has left behind here. See you around.
Mon
Tues
Wed
Thu
FRIDAY
Sat
Sun
October 18, 2002 - Updates at 0705
Good morning. I'm running late, so you get Sally pictures. She's doing well, a little perkier and more mobile each day. By the time six weeks of casts and such are done with, I think we're all going to be a little crazy. Not to mention just how much I like shoving glucosamine down her throat four times a day. That'll moderate in a while, too. But for now, she can just hate it. After all, the rest of the time, she gets lots of lovin' from us. And, as you can see, she still manages to find a way to curl up in her bed.
Another one from the Ooooooh-Isn't-That-Brilliant department: The consul in Moscow is refusing to grant visas to Dmitry Sklyarov and Alexander Katalov of ElcomSoft. They're trying to return here as defendents in the DMCA case that Adobe got rolling a while back. I'm wondering if this isn't really a government ploy to get the charges dropped so that the DMCA doesn't come up in a case where it might actually lose? The boys are supposed to be here next Thursday, so schedules are getting a little tight.
Yesterday morning, I was finally successful in penetrating the download queues and started to fetch down copies of United Linux, Beta 3. I've done an initial installation in VMware (yay, at least that works for THIS distro), and let me tell you. I don't know why anyone but SuSE is part of this gang. It's YAST that handles installation and configuration. So what's left that's personal to a specific vendor beyond the installer and config tools? Nothing, to my mind. I'll do another round, with screenshots and commentary, for LinuxMuse right soon.
Gotta roll. Have a great day!
Mon
Tues
Wed
Thu
Fri
SATURDAY
Sun
October 19, 2002 - Updates at 1150
Good morning. I've been a busy beaver already today. After sleeping in just a bit, I got the call to head into the office and hard-reset one of our Raq3 servers which had thoughtlessly taken itself offline. One pair of power switch state changes later, the box was back online and in service. Since then, I've had a late spot of breakfast, my morning pharmacopoeia (comprised these days of an AntiOxidant, Glucosamine + MSM, Centrum Zinc to A, and Ginseng). and plowed through emails that have been piling up since yesterday morning.
Yes, since yesterday morning. I had a long, yet productive day yesterday at the standard Friday client site. I rebuilt the storage server into a configuration that makes a bit more sense. Here's the hardware: A 2U rack server with Intel mobo, PIII-933 and 128M RAM. A Promise UltraATA-133 add-in IDE card, and four (4) Maxtor 120 G drives. Our first configuration involved a 3 G root partition, a 1 G swap, and the balance of all 4 drives joined in a single LVM volume of 464 G to serve as a backup staging server. While perhaps not the most stable configuration on the block, it would have done fine for transient data only. That is, production servers backup across the net to the backup staging server at net speed, then backups to tape happen from the staging server at tape speed, reducing the backup load time for the production boxes.
Then the decision was made a week ago to move the assorted NFS and SMB mounted homes off of a Snap Server and onto the staging server as well. At that point I expressed my discomfort with keeping production data available on a relatively untested configuration where the failure of any one of four spindles breaks the whole setup. So yesterday, I rebuilt the box from scratch. Drives one and two are root, swap, and a 217 G LVM volume. The latter stands in use for backup staging. Drives three and four are software RAID 1, mirrored and providing 117 G of solid storage for the homes. Everything was spinning like a top, and we'd started rsyncing the data over to the new space before I left for the day.
In between phases of that setup, I also revised another Linux box to serve as a specialized single service gateway between two different networks. On the outside network, the server NFS mounts a set of data from another machine. On the inside, Samba is configured to share that NFS mount read-only. Thus we don't have to open a SMB hole through the firewall, and all services are gated by IPtables, with specific machine/IPs trusted and all others denied. Sweet setup, and one that we didn't think I'd have up and running before the end of the day yesterday. Woo Hoo!
Today, I'm mucking about a bit with United Linux Beta 3, a Neverwinter Nights server, and who knows what all else. Mmmm. Almost lunchtime, too. See you around!
Mon
Tues
Wed
Thu
Fri
Sat
SUNDAY
October 20, 2002 - Updates at 0950
Another right coast morning, another probable sniper victim. And this guy's not taking weekends off anymore, eh? What a weird world we live in. Good morning. It's a gray day here, with light rain likely tonight into tomorrow morning. Marcia and I today are tackling household chores and other stunningly exciting things like grocery shopping.
Yesterday's experiments running the Neverwinter Nights Linux Server were successful. I was running it on the hardware once known as Grendel, a P2-233 with 256M of RAM. It might struggle with tens or hundreds of simultaneous user connections, but for Greg and I on a small campaign, it rocked. I don't have stats yet, but I'll track the system load as we keep testing.
That's all I've got for now, maybe more after another half-cup of coffee. Enjoy!
Mon
Tues
Wed
Thu
Fri
Sat
Sun
Last Week
<-- * -->
Next Week
Visit the rest of the DAYNOTES GANG, a collection of bright minds and sharp wits. Really, I don't know why they tolerate me <grin>. My personal inspiration for these pages is Dr. Jerry Pournelle. I am also indebted to Bob Thompson and Tom Syroid for their patience, guidance and feedback. Of course, I am sustained by and beholden to my lovely wife, Marcia. You can find her online too, at http://www.dutchgirl.net/. Thanks for dropping by.
All Content Copyright © 1999-2002 Brian P. Bilbrey.