Dima's questions

I'm trying to setup TFTP server on CentOS 6.2. The /etc/xinet.d/tftp configuration file is the following:

service tftp
{
    disable = no
    socket_type     = dgram
    protocol        = udp
    wait            = yes
    user            = root
    server          = /usr/sbin/in.tftpd
    server_args     = -s /tftpboot -vvv
    per_source      = 11
    cps         = 100 2
    flags           = IPv4
}

The selinux and firewall are disabled. The /etc/hosts.allow and /etc/hosts.deny files are empty. When I'm trying to get a file from the TFTP server, the file transfer always failed and I see the following errors into /var/log/messages

Jul 11 03:16:53 localhost xinetd[4155]: xinetd Version 2.3.14 started with libwrap loadavg labeled-networking options compiled in.
Jul 11 03:16:53 localhost xinetd[4155]: Started working: 1 available service
Jul 11 03:17:00 localhost xinetd[4155]: START: tftp pid=4157 from=192.168.10.3
Jul 11 03:17:00 localhost in.tftpd[4158]: RRQ from 192.168.10.3 filename 1
Jul 11 03:17:00 localhost in.tftpd[4158]: sending NAK (0, Permission denied) to 192.168.10.3
Jul 11 03:17:01 localhost in.tftpd[4159]: RRQ from 192.168.10.3 filename 1
Jul 11 03:17:01 localhost in.tftpd[4159]: sending NAK (0, Permission denied) to 192.168.10.3
Jul 11 03:17:03 localhost in.tftpd[4160]: RRQ from 192.168.10.3 filename 1

The tftpboot directory permissions are (output of the ls -l command):

drw-rw-rw-.   3 root root  4096 Jul 11 03:32 tftpboot

I also see that the tftpboot directory is shown (by ls -l) with green background (unlike other files/directories) (Why? As I know the green background is for sticky bit only).

What I did wrong? How can I make TFTP server working?

I'm trying to compile and install kernel 3.2.9 for CentOS 5.6. I compiled new kernel and installed kernel modules:

make && make modules_install

I also created new initrd image:

mkinitrd initrd-3.2.9 3.2.9

I copied kernel-3.2.9, initrd-3.2.9 and System.map-3.2.9 to the /boot directory and edited grub.conf file:

title kernel-3.2.9
        root (hd0,0)
        kernel /kernel-3.2.9 ro root=LABEL=/ rhgb quiet
        initrd /initrd-3.2.9

During boot I got a kernel panic with the following error:

could not find filesystem /dev/root

What I did wrong?

I need to route a multicast streams in my private network. I'm using pimd for multicast routing. The multicast routing did not work until I put altnet <multicast_source_ip> masklen 24 to the pimd.conf file. The problem that I do not always know what the <multicast_source_ip>.

I still not understand what the altnet directive is doing? According to the pimd.conf file:

# If you want to add "alternative (sub)net" to a physical interface,
# e.g., if you want to make incoming traffic with a non-local source address
# to appear as it is coming from a local subnet, then use the command:
#   phyint <local-addr | ifname> altnet <net-addr> masklen <len>
# XXX: if you use this command, make sure you know what you are doing!!

And how can I solve this problem without <multicast_source_ip> using?

I'm trying to pass mulitcast stream from interface eth1 (192.168.20.41) to interface tun0 (192.168.100.40) on CentOS 5.

I can see incoming multicast stream on eth1:

tcpdump -n -i eth1

type=1700 audit(1324681169.542:52): dev=eth1 prom=256 old_prom=0 auid=4294967295 ses=4294967295
Dec 23 17:59:29 localhost kernel: device eth1 entered promiscuous mode
Dec 23 17:59:29 localhost kernel: type=1700 audit(1324681169.542:52): dev=eth1 prom=256 old_prom=0 auid=4294967295 ses=4294967295
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth1, link-type EN10MB (Ethernet), capture size 96 bytes
17:59:29.576192 IP 192.168.20.20.52194 > 224.1.1.1.search-agent: UDP, length 1328
17:59:29.576277 IP 192.168.20.20.52194 > 224.1.1.1.search-agent: UDP, length 1328
17:59:29.576801 IP 192.168.20.20.52194 > 224.1.1.1.search-agent: UDP, length 1328

But I cannot see multicast on tun0 interface. What I'm doing wrong? The configuration is attached below:

/etc/igmpproxy.conf file:

phyint eth1 upstream  ratelimit 0  threshold 1
        altnet 192.168.100.0/24
phyint tun0 downstream  ratelimit 0  threshold 1
phyint eth0 disabled
phyint eth5 disabled

iptable configuration:

iptables -A INPUT -p igmp -j ACCEPT
iptables -A INPUT -d 224.0.0.0/240.0.0.0 -p udp -m udp -j ACCEPT
iptables -A FORWARD -d 224.0.0.0/240.0.0.0 -p udp -j ACCEPT
modprobe ipt_TTL
iptables -t mangle -A PREROUTING -d 224.0.0.0/240.0.0.0 -p udp -j TTL --ttl-inc 1

tun0 adapter is GRE tunnel over eth0: cat /etc/sysconfig/network-scripts/ifcfg-tun0

DEVICE=tun0
TYPE=GRE
ONBOOT=yes
MY_INNER_IPADDR=192.168.100.40
PEER_INNER_IPADDR=192.168.100.30
PEER_OUTER_IPADDR=192.168.20.30

I need to drop UDP packets with specific UDP port number) on network interface (say eth0). On my setup the interface eth0 is part of the bridge.

is it possible to do it using tc command? if yes, how?

I have an disk image with 3 partitions: first partition (hd0,0) is boot-able with GRUB1 with the following configuration GRUB file:

default=0
timeout=5
title Bank A
        root (hd0,1)
        chainloader +1

title Bank B
        root (hd0,2)
        chainloader +1

The partitions (hd0,1) and (hd0,2) are also boot-able. I'm trying to clone partition (hd0,1) to (hd0,2) by creating device map using kpartx and copying whole partition using dd command.

The problem is: after partition cloning, the cloned partition did not boot (but all files are OK). What the wrong?

I need both partitions to bee identical (I'm using them for fail-over purposes into embedded device)

I have disk image with 2 partitions:

1. Partition 1 has cramfs file system (read only). This partition contains all system files of the OS
2. Partition 2 has ext3 file system. This partition has only configuration files that may be changed.

How can I install GRUB1 boot loader on MBR.

I tried to copy first 446 bytes of my hard disk and copy GRUB files to the /boot directory on the 1st (cramfs) partition.

I cannot use grub-install because I have disk image and not disk itself.

Any ideas?

I'm trying to create disk image with 2 partitions: First partition: 1Mb for boot and second partition of the maximum capacity

  dd if=/dev/zero of=flash.img bs=1M count=10
  /sbin/sfdisk flash.img -uM << SFDISK_EOF
,1,L,*
,,L;
SFDISK_EOF

I got an error:

sfdisk: I don't like these partitions - nothing changed.

What I'm doing wrong?

I'm working with CentOS

I have 3 images. Each image was done by cloning (by dd) a partition (these partitions formatted using ext3 file system). One partition is boot-able with GRUB1 (CentOS)

How can I combine them to one disk image?

(I can copy the partitions images using dd command, but I don't know how to create partitions table and make boot working).

I need to install CentOS 5 server using kickstart installation. As part of the installation, I need to install lm_sensors package. When I'm installing lm_sensors manually, I'm running sensors-detect command. But this command is requires an interaction with user.

What I need to do in order to make fully automatic CentOS installation with lm_sensors?

Thanks

Where can I download precompiled CentOS 5 kernel including realtime patch? Or I need to compile it myself (if yes, how to do that)?

Can I know if my service is started on system init or it started by user using service command?

On my service I need to make different logic (into service starter script) if service is started automatically or by user command.

Thanks

I need to add an additional title to my grub.conf file and after OS boot I need start different services depends on it

What the better way to do that?

I need to run some bash script only once on first boot (after OS installation) of my CentOS machine This script should run before network service start (because this script will change network configuration) How can I register my script to run before network service start?

Today, in order to install CentOS 5.5 I'm using kickstart script.

I would like to install CentOS on different way:

1. Create disk image (using dd command)
2. Create filesystem on this disk image using mkfs.ext3
3. Install CentOS on this filesystem
4. Make this disk image bootable (using grub-install)
5. Copy the disk image to the physical hard disk (using dd command)

I know to do all these items except 3. Is it possible to do it? If yes, how can I install CentOS on the disk image?