Note: This web page was automatically created from a PalmOS "pedit32" memo.

Fedora Core 3 notes (FC3)

I did almost an entire install, read the release notes, and took notes, through PalmVNC. However, it ran out of juice part way through due to the heavy 802.11b usage, so I had to use it from its cradle part of the time.
If you want to verify your install media during an install, make sure to boot with "linux ide=nodma". Otherwise, odds are your first CD will verify, but the others won't. Alternatively, you may be able to burn your install CD's with cdrecord -pad. I didn't monitor the time estimates provided by anaconda closely, but I noticed that sometimes the estimate would go up. Burning with padding -does- appear to help matters. You get it with "cdrecord -pad ..."
The install took a LONG time. I wish I'd timed it. I started the install around breakfast time, and it was still going after my late lunch. This was on a 56x CD-ROM drive.
Timezone pick died clicking on map - had to restart the install. Suggest using City/TZ list instead.
install (anaconda) uses LVM now for filesystems it creates.
install has useful-sounding warn option for selinux: Nothing is denied, but you are told what would have been denied.
installing everything took ~6.7 gigabytes.
release notes came up on second click. First click generated an error.
probably want export MOZ_ENABLE_PANGO=1, for both mozilla and firefox.
ssh -X is more cautious than it used to be. If you want the old behavior, use ssh -Y.
ext2online can resize ext3. It does Not resize the block device; you'll have to do that another way (fdisk magic, LVM, md, whatever).
MALLOC_CHECK_ (env var) enabled by default in glibc. Some programs may not run as a result. Different values in this env var control how strict the checking is, if any.
If you install a new kernel, it will be made the default kernel. Controlled by /etc/sysconfig/kernel
Foreign language input: iiimf gimlet ctrl-space and shift-space toggle on and off Setting a suitable locale may help, EG zh_CN
gok (gnome onscreen keyboard) may help with palmvnc. However, gok requires the XKB X server extension, and RealVNC 4.0 does not appear to support it. What about TightVNC?
nscd may return items from its cache even after their TTL has expired This may mask name service problems
udev: Check out /etc/udev. You can set permissions on device files here, among other things.
Your first "yum update" after the install tells you to: rpm --import public.gpg.key ...however, this does not work. Instead, you must: [root@tesuji ~]# rpm --import /usr/share/rhn/RPM-GPG-KEY [root@tesuji ~]# rpm --import /usr/share/rhn/RPM-GPG-KEY-fedora ...and then: [root@tesuji ~]# yum -y update
system-config-authentication was all that was required to get both NIS and automounting going.
Had to symlink /bin/bash to /bin/bash2 in order to log in with my strombrg account, due to /dcs/bin/bash2 symlink.
Evolution, upon first invocation, migrated my evo 1.x data to 2.x format.
Had to run system-config-date to get the correct time set.
I did the install from 4 cdroms. I stuck around to change 1-2, 2-3, and 3-4, but as soon as I put in the 4th cdrom, Ieft for home. Happily, despite the 4th cdrom being left in the drive, and the BIOS being set to boot from cd before a hard disk, I was still able to reboot into FC3 on the hard disk from home, having connected again via VNC, this time only to click the "reboot" button.
yum update took a -very- long time. Hours. I chose to use netstat -ap to get a list of pid's of processes that were listening for incoming traffic, and killed each of them, hoping to deter any sort security problems that might occur before the patching could complete. Another alternative is to boot singleuser, run /etc/rc2.d/S10network start, and do your first yum -y update from singleuser.
The system is prelinking a massive number of executables. This appears to be contributing substantially to the slowdown.
As far as jpilot, so far I have been making /etc/udev/devices/ttyUSB1 creating a device node in a place that's supposed to make the device file show up in /dev, but doesn't. So instead, I'm using the path to the source device node. :) Command to create it: mknod /etc/udev/devices/ttyUSB1 c 188 1 chown strombrg /etc/udev/devices/ttyUSB1
One of my FC3 systems has begun crashing a lot. At first I thought it was just when backing up my palm pilot, but after the second reboot, it began hanging when trying to add swap space during boot. Strangely, when I went to reinstall, I didn't see a place to request a bad block scan... Trying to re-mkswap... This appears to have made the system bootable again - but for how long?
Must try writing cdroms with padding to see if that can speed up the install.
The USB support in FC3 may be a bit rocky. I've seen crashes, and devices that don't show up on the bus at all, that worked fine with FC2.
I've been working around a Palm USB sync crash (which happens on one machine and not another) by using netsync instead of USB sync. It's all tunneled. :)
o yum and yum20 [atrpms] name=Fedora Core $releasever - $basearch - ATrpms baseurl=$releasever/en/$basearch/at-stable
Fedora FAQ:
I backed up an FC2 system to a Tru64 host into an NFS volume mounted from a Solaris 2 host, over ssh, gzip'd. Upon attempting to extract data from this archive after upgrading the system to FC3, I found that I couldn't get the data back, using the tar and gunzip that come with FC3. However, once I sftp'd an older gnu tar binary from another system, I could then extract the data on FC3.
Mark Flanner (ESS grad student) had trouble with FC3, right after the install, getting stuck at "loading kernel parameters". He says the problem went away when he did something with the nvidia (graphics) driver.
cdrecord -v -pad -dev ATAPI:0,0,0 foo.iso
> and dump the device-mapper status: > dmsetup info -c > dmsetup status > dmsetup table
I finally got annoyed enough with that "X tunneling doesn't work with an FC[23] ssh server -sometimes-" problem to track down its source. It appears that sometimes FC3's (and probably FC2's) sshd will only attempt to set up X11 tunneling on IPV6, instead of both IPV[46]. This seems to work around it well enough for the time being: 1) Put "OPTIONS=-4" in /etc/sysconfig/sshd 2) Run "service sshd restart" A more long-term solution -might- be to rebuild sshd with "--with-4in6".
SMC 2802W has a Prism GT chipset, and is can be ordered for next-day, in store pickup. 2005-01-22 04:09:41 pm. It's an 802.11b/g card for a PCI bus. 2005-01-31 10:35:28 am It wasn't in the store. I wound up getting a Hawking 802.11b/g card because a list on the web said it had a prism chipset - but it didn't. In fact, it appears that this same "model" of card has had three different chipsets! There's a native linux driver for the chipset my card has, but it didn't work. I wound up using ndiswrapper.
nsrexecd required a change in /etc/hosts? Appears I had to remove seki from the loopback address, and put in a line with seki's real address.
For diff'ing, tweaking rpm's:
If you have an FC3 system that has strange high load, no hungry processes, no large memory usage, try adding this to grub.conf on your kernel lines: pci=routeirq
This seems to work better on tesuji. The burnfree part appears to be important to avoid burning coasters. cdrecord -v -driveropts=burnfree -pad -dev ATAPI:1,0,0 FC3-i386-disc1.iso
If you boot with: linux askmethod vnc vncpassword=whatever ...and specify and /mirrors/fedora/linux/core/3/i386/os the ftp server to install from, then you really only need the first CD.
One of my FC3 systems got royally messed up by using the ATrpms repository. I hunted around the net for solutions, but eventually would up reinstalling. I found a couple of suggested solutions, like upgrading yum, or using apt-get instead, but neither sorted out the mess.
> # cat /etc/udev/rules.d/10-udev.rules > KERNEL="ttyUSB1",SYMLINK="pilot" My entry in /etc/udev/rules.d/50-udev.rules KERNEL="ttyUSB[0-9]*",SYMLINK="pilot" BUS="usb", KERNEL="ttyUSB[0-9]*", NAME="usb/%k" It may help for you to know that what you placed in your file wouldn't work for me. How I came across the syntax above is by looking at the entry for my Epson CX-5400, which was found, driven and worked from the onset. This was the *only* syntax that worked for me. > # cat /etc/udev/permissions.d/10-udev.permissions > #set Palm Pilot rwx > pilot*:root:usb:0666 I have this: # pilot/palm devices pilot:root:uucp:0660 This was 'stock'. I didn't add it. When I saw this, I made sure that my user account had been made a member of 'uucp'. I normally do this from past issues, anyway. /\/\/\/\/\/\/\/\/\/\/\ I found this info to be quite useful for udev and creating entries for unknown' hardware. It enabled me to simply identify my Kyocera 7135 and point it to /dev/pilot and it obviously works because J-Pilot and pilot-xfer programs find it quite easily. gpilotd however - even when told to connect to /dev/pilot remains deaf. That is my point.
Another example of setting up a palm pilot with udev: /etc/udev/rules.d/10-visor.rules BUS="usb", SYSFS{product}="palmOne Handheld*", KERNEL="ttyUSB*", SYMLINK="pilot"
Harry Mangalam recommends "guarddog" for an easy way of configuring iptables firewalls
seki-root> fdisk -l /dev/hda Note: sector size is 2048 (not 512) Disk /dev/hda (Sun disk label): 1 heads, 640 sectors, 2048 cylinders Units = cylinders of 640 * 512 bytes Device Flag Start End Blocks Id System /dev/hda1 r 0 752 240640 4 SunOS usr /dev/hda2 r 752 1984 394240 2 SunOS root /dev/hda3 1984 1988 1280 0 Empty /dev/hda4 1988 1992 1280 0 Empty /dev/hda5 1992 1996 1280 0 Empty /dev/hda6 1996 2000 1280 0 Empty Thu Feb 17 15:41:53 seki-root> b c bash2: b: command not found Thu Feb 17 15:41:55 seki-root> bc bc 1.06 Copyright 1991-1994, 1997, 1998, 2000 Free Software Foundation, Inc. This is free software with ABSOLUTELY NO WARRANTY. For details type `warranty'. 640*512 327680 327680*752 246415360 Thu Feb 17 15:42:13 seki-root> mount /dev/hda /mnt/root Thu Feb 17 15:42:19 seki-root> bc Thu Feb 17 15:42:20 seki-root> mount -o offset=246415360 /dev/hda /mnt/usr Thu Feb 17 15:42:28 seki-root> cd /mnt/usr Thu Feb 17 15:42:31 seki-root> d ./ bin@ devices/ kernel/ opt/ reconfigure .swappart@ .tmp_proto/ ../ cdrom/ etc/ lib/ platform/ sbin/ system/ usr/ a/ dev/ .java@ mnt/ proc/ .swapinfo@ tmp/ var@ /mnt/usr Thu Feb 17 15:42:32 seki-root>
Apparently the Fedora installer set me up with: # Firewall configuration written by system-config-securitylevel # Manual customization of this file is not recommended. *filter :INPUT ACCEPT [0:0] :FORWARD ACCEPT [0:0] :OUTPUT ACCEPT [0:0] :RH-Firewall-1-INPUT - [0:0] -A INPUT -j RH-Firewall-1-INPUT -A FORWARD -j RH-Firewall-1-INPUT -A RH-Firewall-1-INPUT -i lo -j ACCEPT -A RH-Firewall-1-INPUT -i eth0 -j ACCEPT COMMIT However, this works much better: # Firewall configuration written by system-config-securitylevel # Manual customization of this file is not recommended. *filter :INPUT ACCEPT [0:0] :FORWARD ACCEPT [0:0] :OUTPUT ACCEPT [0:0] :RH-Firewall-1-INPUT - [0:0] -A INPUT -j RH-Firewall-1-INPUT -A FORWARD -j RH-Firewall-1-INPUT -A RH-Firewall-1-INPUT -i lo -j ACCEPT -A RH-Firewall-1-INPUT -i eth0 -j ACCEPT -A RH-Firewall-1-INPUT -p icmp --icmp-type any -j ACCEPT COMMIT Please note the addition of the icmp line. Without it, ssh sessions, zebedee sessions, and other protocols work fine for a while, but freeze up solid before very long.
> > 10.1. Is my box posessed? > > In the past I've seen that happen when I switch between boxes with a > KVM. X just freaks out exactly as you described. Using the IMPS/2 > pointer driver in XF86Config fixed the problem. > I found another way to fix kvm switch problems like that. Just put this in your kernel boot options: psmouse.proto=bare This will disable your mouse scroll wheel, but stops the erratic behavior. I rarely use the scroll wheel, so this has been a good fix for me, when I have to work with cheap kvm switches.
udev and palms again:
This finally cured my "yum is horribly broken problem": rpm --erase python-devel Pyrex sip-devel PyQt-devel After that command, "yum -y update" worked like a charm.
/usr/lib/yp/ypinit -m has been designated the Official Community Support Forum of the Fedora Project. This decision has been made by comittee because of the many advantages web forums have over mailing lists in support of end-users. Nothing changes about the existing mailing lists, but we encourage end-users to go to the forums first if they have support questions. Development discussion remains unchanged on designated mailing lists. End-User Official Community Support Forum of Fedora Project All users of Fedora Core and Fedora Legacy are encouraged to join the forums here. Post questions in the appropriate forum, and any threads that you participate can notify you via e-mail when someone replies. It is generally very easy to skip over threads that you don't care about, and find those that you do care.

Back to Dan's palm memos