flashnode's questions

I had to downgrade a couple of Cisco 3925 routers last week and one of the steps in the instructions (link provided) was to change the boot command that was given by the following command:

Cisco-3925# show run | include boot

The downgrade went fine, but what I'd like to know is why this command didn't show any boot instruction before the downgrade was successful. Here is the output of that command:

boot-start-marker
boot-end-marker

How did the router boot before the downgrade in the first place?

Running ltrace isn't trivial. This RHEL 5.3 system has based on a Tomcat Catalina (servlet container) which uses text scripts to tie everything together. When I tried to find an executable here's the rabbit hole I went down:

/etc/init.d/pki-ca9 calls dtomcat5-pki-ca9

#Path to the tomcat launch script (direct don't use wrapper)
TOMCAT_SCRIPT=/usr/bin/dtomcat5-pki-ca9

/usr/bin/dtomcat5-pki-ca9 calls a watchdog program

/usr/bin/nuxwdog -f $FNAME

I replaced nuxwdog with a wrapper

[root@qantas]# cat /usr/bin/nuxwdog
#!/bin/bash
ltrace -e open -o /tmp/ltrace.$(date +%s) /usr/bin/nuxwdog.bak $@

[root@qantas]# service pki-ca9 start
Starting pki-ca9:              [  OK  ]

[root@qantas]# cat /tmp/ltrace.1295036985
+++ exited (status 1) +++

This is ugly. How do I run strace or ltrace in tomcat?

UPDATE

Here is the tomcat "process"

[root@qantas]# ps -ef | grep tomcat
pkiuser  21767 21766  0 10:10 ?        00:00:09 /usr/lib/jvm/jre/bin/java
 -Djava.endorsed.dirs=/usr/share/tomcat5/common/endorsed
 -classpath :/usr/lib/jvm/jre/lib/rt.jar:/usr/share/java/commons-
collections.jar:/usr/share/tomcat5/bin/bootstrap.jar:/usr/share/tomcat5/bin/commons-
logging-api.jar:/usr/share/java/mx4j/mx4j-impl.jar:/usr/share/java/mx4j/mx4j-
jmx.jar:/usr/share/tomcat5/common/lib/nuxwdog.jar -Dcatalina.base=/var/lib/pki-ca11
 -Dcatalina.home=/usr/share/tomcat5 -Djava.io.tmpdir=/usr/share/tomcat5/temp
 org.apache.catalina.startup.Bootstrap start

Here are all the children

[root@qantas]# pstree -A -p 21767
java(21767)-+-{java}(21768)
            |-{java}(21769)
            |-{java}(21770)
            <..snip..>
            `-{java}(22104)

This is what happens when I try to start ltrace on the text script

[root@qantas]# ltrace /usr/bin/dtomcat5-pki-ca11
ltrace: Can't open ELF file "/usr/bin/dtomcat5-pki-ca11"

When I attach ltrace to the parent and let it run for a day there is no output

[root@qantas]# ltrace -e open -o /tmp/ltrace.1295465058 -p 21767

...24 hours later...

[root@qantas]# ls -l /tmp/ltrace.1295465058
-rw-r--r-- 1 root root 0 Jan 19 11:24 /tmp/ltrace.1295465058

I need to determine whether a library named libunaSA.so is being called directly by the process or by another library called libtoki2.so. I guess what I'm looking for is a pstree for libraries. The system is running RHEL 5.3 Beta.

This output shows the two libraries in the process map

# grep -e toki -e una /proc/2335/maps
0043f000-004ad000 r-xp 00000000 08:02 543465     /usr/lib/libtoki2.so
004ad000-004c5000 rwxp 0006d000 08:02 543465     /usr/lib/libtoki2.so
01185000-01397000 r-xp 00000000 08:02 543503     /usr/lib/libunaSA.so
01397000-013dc000 rwxp 00211000 08:02 543503     /usr/lib/libunaSA.so

This output shows that only the libtoki2.so library is in the current cache

# ldconfig -p | grep -e una -e toki
libtoki2.so (libc6) => /usr/lib/libtoki2.so
libtoki.so.4.4.1 (libc6) => /usr/lib/libtoki.so.4.4.1
libtoki.so.2 (libc6) => /usr/lib/libtoki.so.2

I attached strace to the running process but it doesn't provide much output

# strace -p 2335
Process 2335 attached - interrupt to quit
futex(0xb7ef5bd8, FUTEX_WAIT, 2336, NULL

Here's the output to ldd for each library

# ldd /usr/lib/libtoki2.so
        linux-gate.so.1 =>  (0x00a0a000)
        libdl.so.2 => /lib/libdl.so.2 (0x001bd000)
        libstdc++-libc6.2-2.so.3 => /usr/lib/libstdc++-libc6.2-2.so.3 (0x00f3f000)
        libm.so.6 => /lib/libm.so.6 (0x00b27000)
        libc.so.6 => /lib/libc.so.6 (0x0043d000)
        /lib/ld-linux.so.2 (0x00742000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x00110000)
# ldd /usr/lib/libunaSA.so
        linux-gate.so.1 =>  (0x00244000)
        libpthread.so.0 => /lib/libpthread.so.0 (0x00baf000)
        libdl.so.2 => /lib/libdl.so.2 (0x007fa000)
        libstdc++-libc6.2-2.so.3 => /usr/lib/libstdc++-libc6.2-2.so.3 (0x009ce000)
        libm.so.6 => /lib/libm.so.6 (0x00c96000)
        libc.so.6 => /lib/libc.so.6 (0x004a2000)
        /lib/ld-linux.so.2 (0x00742000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x00a9f000)

I installed a demo for HP Network Automation last week and the network device I am trying to interface with is not listed in the "Supported Devices Matrix"

Apparently there is a way to design a custom driver, but so far all I have been able to find is an API guide. A quick search through the guide makes no mention of customer drivers for polling devices.

Does anyone have experience with HPNA custom drivers and If so what's my first step?

If I can't ssh as root to each of my servers how can I make modifications in an efficient way?

I am not allowed to setup ssh keys or open the sudoers file with NOPASSWD. I can't install puppet or spacewalk.

Sometimes when I try to include a sudo command in a script I get the error "no tty present." Has anyone worked in an environment like this?