bradlis7's questions

I'm setting up a Virtual Machine, and I'm trying to get OCFS to allow redundancy across two devices. It looks to be working in configuration, but it's not actually syncing the files across:

/etc/ocfs2/cluster.conf

cluster:
        node_count = 2
        name = vmcluster

node:
        ip_port = 7777
        ip_address = 198.18.0.25
        number = 1
        name = vm1
        cluster = vmcluster

node:
        ip_port = 7777
        ip_address = 198.18.0.26
        number = 2
        name = vm2
        cluster = vmcluster

/etc/sysconfig/o2cb

# O2CB_ENABLED: 'true' means to load the driver on boot.
O2CB_ENABLED=true

# O2CB_STACK: The name of the cluster stack backing O2CB.
O2CB_STACK=o2cb

# O2CB_BOOTCLUSTER: If not empty, the name of a cluster to start.
O2CB_BOOTCLUSTER=vmcluster

# O2CB_HEARTBEAT_THRESHOLD: Iterations before a node is considered dead.
O2CB_HEARTBEAT_THRESHOLD=

# O2CB_IDLE_TIMEOUT_MS: Time in ms before a network connection is considered dead.
O2CB_IDLE_TIMEOUT_MS=

# O2CB_KEEPALIVE_DELAY_MS: Max time in ms before a keepalive packet is sent
O2CB_KEEPALIVE_DELAY_MS=

# O2CB_RECONNECT_DELAY_MS: Min time in ms between connection attempts
O2CB_RECONNECT_DELAY_MS=

Status of OCFS Programs:

[root@vm1 ~]# /etc/init.d/ocfs2 status
Configured OCFS2 mountpoints:  /cluster1 /cluster2
Active OCFS2 mountpoints:  /cluster1 /cluster2

[root@vm1 ~]# /etc/init.d/o2cb status 
Driver for "configfs": Loaded
Filesystem "configfs": Mounted
Stack glue driver: Loaded
Stack plugin "o2cb": Loaded
Driver for "ocfs2_dlmfs": Loaded
Filesystem "ocfs2_dlmfs": Mounted
Checking O2CB cluster "vmcluster": Online
  Heartbeat dead threshold: 31
  Network idle timeout: 30000
  Network keepalive delay: 2000
  Network reconnect delay: 2000
  Heartbeat mode: Local
Checking O2CB heartbeat: Active

I see that the heartbeat mode says "Local". I've researched the documentation at http://oss.oracle.com/projects/ocfs2/dist/documentation/v1.6/ocfs2-1_6-usersguide.pdf and old tutorials at http://www.hbyconsultancy.com/blog/shared-ocfs2-partition-on-ubuntu-server-10-04-x64.html and other places. I am not seeing the whole picture it seems...

I'm using Oracle Linux 3.0.1, which is a modified version of RHEL6.

Our company has a server that is a locked down version of Linux (custom shell for users, no direct access). I am currently pushing updates to the user, but it would be more ideal for the users to do a pull.

Each device has a serial number that would probably be used as part of the login to an update server.

I would like to have multiple channels/branches (whatever they should be called), such as MAIN and DEV.

What protocol would be best for this situation? FTP is easier to limit directory access, in case someone was to get the full login information. SCP would be better, as I would like to secure the transmission, but I would prefer read-only and limited to a directory. Can this be done with SCP?

What should the username/password be? I guess for SCP, the serial number could be the username, and instead of a password, generate a key for the user.

I'm trying to set up a firewall with support for snort, and it is dropping all of my packets when I add the QUEUE target. I've made it like this, but the QUEUE target is not allowing the packets to be processed any further:

-A INPUT -p tcp -m tcp --dport 22 -j ACCEPT 
-A INPUT -j QUEUE 
-A INPUT -j ACCEPT

# It's not allowing anything past QUEUE, as you can see below in the count.
> iptables -I INPUT -nv
 pkts bytes target     prot opt in     out     source               destination         
 6707  395K ACCEPT     tcp  --  *      *       0.0.0.0/0            0.0.0.0/0           tcp dpt:22 
  933  138K QUEUE      all  --  *      *       0.0.0.0/0            0.0.0.0/0           
    0     0 ACCEPT     all  --  *      *       0.0.0.0/0            0.0.0.0/0           

I'm eventually going to change it to forward, but I'm just trying to get it working for now. I start snort like so:

snort -Q -D -c /etc/snort/snort.conf

EDIT: More Information

When I run it, it still sees the packets without having an iptables QUEUE target rule, but when I add a QUEUE target, it starts losing all of my packets.

# snort -Qc /etc/snort/snort.conf -N -A console
Enabling inline operation
Running in IDS mode

        --== Initializing Snort ==--
Initializing Output Plugins!
Initializing Preprocessors!
Initializing Plug-ins!
Parsing Rules file "/etc/snort/snort.conf"
## === CUT ===
***
*** interface device lookup found: bond0
***
Initializing Network Interface bond0
Decoding Ethernet on interface bond0
## === CUT ===
Not Using PCAP_FRAMES

So, it says inline, but the it says it's using bond0. Inline should not require an interface, right?

I have set up a test ISC DHCP server that is not connected to an internet with DHCP and DHCPv6. I also have a Cisco 3750G with IPv6 enabled. I have firewalls off on both machines.

My Windows 7 machine will get both an IPv4 and an IPv6 address:

Ethernet adapter Wired Network Connection:

   Connection-specific DNS Suffix  . :
   Description . . . . . . . . . . . : Intel(R) 82566MM Gigabit Network Connection
   Physical Address. . . . . . . . . : 00-0B-97-DD-18-4E
   DHCP Enabled. . . . . . . . . . . : Yes
   Autoconfiguration Enabled . . . . : Yes
   IPv6 Address. . . . . . . . . . . : 2001:6::afb(Preferred)
   Link-local IPv6 Address . . . . . : fe80::900c:ec5b:cfd0:8470%11(Preferred)
   IPv4 Address. . . . . . . . . . . : 10.99.1.101(Preferred)
   Subnet Mask . . . . . . . . . . . : 255.255.0.0
   Default Gateway . . . . . . . . . :
   DHCP Server . . . . . . . . . . . : 10.99.1.10
   DHCPv6 IAID . . . . . . . . . . . : 234883991
   DHCPv6 Client DUID. . . . . . . . : 00-01-00-01-13-99-C3-BE-00-0B-97-DD-18-4E

   DNS Servers . . . . . . . . . . . : 2001:6::10
                                       10.99.1.10
   NetBIOS over Tcpip. . . . . . . . : Enabled

Pinging on IPv4 works fine both ways. I can ping using IPv6 from the DHCP server to my windows machine, but if I try to ping from the Windows machine, to elsewhere, it gives me this:

C:\Users\Brad>ping 2001:8::10

Pinging 2001:8::10 with 32 bytes of data:
PING: transmit failed. General failure.
PING: transmit failed. General failure.

Ping statistics for 2001:8::10
    Packets: Sent = 2, Received = 0, Lost = 2 (100% loss)

I read that I need to have a default router set (doesn't really make sense as I'm not leaving the network), but I can't find the option to set it. option routers expects an IPv4 address.

Any ideas on what I'm doing wrong?

Edit: More Info

/etc/dhcpd6.conf

#Some defines
# These made no difference, but I saw them elsewhere:
#option dhcp6.softwire code 54 = ip6-address;
#option dhcp6.defroute code 123 = ip6-address;
#option dhcp6.defgateway code 99 = ip6-address;

ping-check false;
ddns-update-style none;
authoritative;
default-lease-time 900;
min-lease-time 900;
max-lease-time 900;

shared-network "pool 1" {
  subnet6  2001:6::/64 {
    range6 2001:6::11 2001:6::fff;
    #prefix6 2001:6::1:0:0:0:0 2001:6::ffff:0:0:0:0 / 64;
    option dhcp6.name-servers 2001:6::10;
    option dhcp6.defgateway 2001:6::1;
  }
}

Windows 7 Routing Table (It's only giving me /128 for the IPv6 address)

IPv6 Route Table
===========================================================================
Active Routes:
 If Metric Network Destination      Gateway
  1    306 ::1/128                  On-link
 11    266 2001:6::afb/128          On-link
 33    276 fe80::/64                On-link
 14    281 fe80::/64                On-link
 11    266 fe80::/64                On-link
 14    281 fe80::39e6:c23:1fc4:b71b/128
                                    On-link
 11    266 fe80::900c:ec5b:cfd0:8470/128
                                    On-link
 33    276 fe80::b933:ca7f:14:f58d/128
                                    On-link
  1    306 ff00::/8                 On-link
 33    276 ff00::/8                 On-link
 14    281 ff00::/8                 On-link
 11    266 ff00::/8                 On-link
===========================================================================
Persistent Routes:
  None

I am trying to monitor for network failures on a machine, and one portion of that is to monitor interfaces that are intended to be active also be "RUNNING".

An Ethernet port, such as eth0, will say "RUNNING" if it is physically connected to another device.

The problem lies in the bonded interfaces, such as bond0. If all of the ethernet devices are disconnected, it still says that it is running, and it is still pingable. Is this by design, or is my system setup incorrectly? Does the miimon option have something to do with this?

EDIT

I found a possible issue, thanks to Gregor. When I ran cat /proc/net/bonding/bond0, it gave me this (notice MII Polling Interval):

Bonding Mode: load balancing (round-robin)
MII Status: up
MII Polling Interval (ms): 0
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth7
MII Status: up
Link Failure Count: 0

But in my /etc/modprobe.conf, I set it up like this:

alias bond0 bonding
options bond0 -o bond0 mode=0 miimon=100

So, evidently, the miimon setting is not working correctly.

This seems like it should be fairly easy to do, but I've run into a few problems.

I've added a cron job to parse all users whose UID is greater than 500:

## This checks if the passwd file has been updated, and generates
## the .forward file from all users whose UID > 500, excluding nobody.
* * * * *       root /usr/bin/test /etc/passwd -nt ~allusers/.forward \
&& /bin/egrep '([5-9]|[0-9]{2})[0-9]{3}' /etc/passwd | /bin/grep -v 65534 \
| /bin/cut -d ':' -f 1 > ~allusers/.forward

Then I created a .procmailrc file:

VERBOSE=yes
LOGFILE=/var/log/procmailrc


#Allow only certain users to send
:0
* ^From.*[email protected].*
{}
:0E
/dev/null

But, the .forward file is processed before it even gets to procmail, evidently. If I moved the .forward file to another filename, can I use it in procmail to send an email to the users in this file?

We are running an outbound SPAM filter, and we'd like to be able to have the mail sent through that filter, even when it's to a local user. Is there a way to force sendmail to use the SMART_HOST, even when it is going to be delivered locally?

Example

Here's what I want. My host is example.com. If I send an email from [email protected] (using SMTP) to [email protected], I want the email to go from mail.example.com to outbound.example.com to inbound.example.com, then back to mail.example.com. Is this possible without creating an endless loop?

I have this code in /etc/procmailrc:

DROPPRIVS=yes

DEFAULT=$HOME/Maildir/

:0
* ? /usr/bin/test -d $DEFAULT || /bin/mkdir $DEFAULT
{ }
:0 E
{
        # Bail out if directory could not be created
        EXITCODE=127
        HOST=bail.out
}

MAILDIR=$HOME/Maildir/

But, when the directory already exists, sometimes it will send a return email with this error: 554 5.3.0 unknown mailer error 127. The email still gets delivered, mind you, but it sends back an error code to the sending user as well.

I fixed this temporarily by commenting out the EXITCODE and HOST lines, but I'd like to know if there is a better solution.

I found this block of code in multiple places across the net, but couldn't really find why this error was coming back to me. It seems to happen when I send an email to a local user. Sometimes the user has a .forward file to send it on to other users, sometimes not, but the result has been the same. I also tried removing DROPPRIVS, just in case it was messing up the forwarding, but it did not seem to affect it.

• Is the line starting with * ? /usr/bin/test a problem?
• The * signifies a regex, but the ? makes it return an integer value, correct?
• What is the integer being matched against? Or is it just comparing the integer return value?
• Do I need a space between the two blocks?

Thanks for the help.

I'm trying to use procmail to send all messages for a certain domain to RT (Request Tracker). This is more of a question about the .procmailrc file, though.

Here's my current .procmailrc file:

#Preliminaries
SHELL=/bin/sh               #Use the Bourne shell (check your path!)
MAILDIR=${HOME}        #First check what your mail directory is!
LOGFILE=${MAILDIR}/procmail.log
LOG="--- Logging ${LOGFILE} for ${LOGNAME}, "
VERBOSE=yes
MAILDOMAIN='rt.mydomain.com'
RT_MAILGATE="/opt/rt3/bin/rt-mailgate"
RT_URL="http://rt.mydomain.com/"

LOGABSTRACT=all

### Trying to  process using the rt-mailgate script
:0
{
# the following line extracts the recipient from Received-headers.
# Simply using the To: does not work, as tickets are often created
# by sending a CC/BCC to RT
TO=`formail -c -xReceived: |grep $MAILDOMAIN |sed -e 's/.*for *<*\(.*\)>* *;.*$/\1/'`
QUEUE=`echo $TO| $HOME/get_queue.pl`
ACTION=`echo $TO| $HOME/get_action.pl`
:0 h b w 
|/usr/bin/perl $RT_MAILGATE --queue $QUEUE --action $ACTION --url $RT_URL
}


### Upon failure, I want to send back an error message to the user, saying
###   "Queue does not exist." I took this code from the procmailex manpage.
:0 Wh: no-queue.lock
{
## Reply if error
* !^FROM_DAEMON
* !^X-Loop: [email protected]
| formail -rD 8192 no-queue.cache

  :0 ehc
  |(formail -rI"Precedence: junk" -A"X-Loop: [email protected]" ; \
    echo "The Queue or Action was invalid."; echo "--" \
    ) | $SENDMAIL -oi -t

}

Do you see a problem with my .procmailrc file? It works fine if the queue exists, but after that point, it just sends the mail to /var/mail/username. I want to throw away the email and return an error message.

I'm doing some research, and trying to find out if any processors do not support PAE that are currently being used. Wikipedia has a short list (Pentium M, Celeron M, Pentium PRO... I found this list under Executable Space Protection ).

Are the processors that do not support PAE being used in current systems? I'm looking at how much risk there is in setting PAE as the default mode in a linux distribution.

Thanks for the help!

Edit: (By the way, I am looking into this for the admins of the Devil-Linux Linux distribution.)