Difference between revisions of "Category:Linux troubleshooting"

Latest revision as of 19:11, 6 March 2015

This category will contain a collection of articles on troubleshooting Linux. It will also include a lot of random commands I use for troubleshooting. It will be highly biased towards Red Hat-based (e.g., CentOS)) and Debian-based (e.g., Ubuntu) distros, but most of the commands should work on most Linux distros. I am also using a 64-bit (x86_64) kernel, so my articles will also be biased towards these systems.

DISCLAIMER: THIS INFORMATION IS PROVIDED TO YOU "AS-IS". NO WARRANTIES OF ANY KIND, EXPRESSED OR IMPLIED, ARE MADE TO YOU AS TO THE CONTENT OF THIS PAGE OR ANY MEDIUM IT MAY BE ON, INCLUDING BUT NOT LIMITED TO WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

I DISCLAIM ALL RESPONSIBILITY FOR THE ACCURACY AND RELIABILITY OF INFORMATION ON THIS PAGE OR PAGES IT LINKS TO.

Emergency reboot

see: wikipedia:Magic SysRq key

Alt + SysRq + REISUB ("Raising Elephants Is So Utterly Boring"; execute in slow succession)

For the above to work, you must have it enabled first:

echo "1" >/proc/sys/kernel/sysrq

In order to have it always enabled (e.g., after a reboot), one must in enable it in the /etc/sysconfig/sysctl file (in SuSE, at least):

ENABLE_SYSRQ="yes"

For further information see /usr/src/linux/Documentation/sysrq.txt

System information

$ dmesg
$ iostat
$ vmstat  # to quickly monitor CPU, memory, and I/O usage and decide which is the bottleneck

$ ps -ef | egrep '^root ' | gawk '{print $2}'  # method 1
$ pgrep -u root                                # method 2

$ cat /proc/cpuinfo
$ cat /proc/partitions
$ cat /proc/meminfo
$ cat /proc/sys/vm/swappiness  # number from 0 - 100; the higher the number the more the system will swap
$ cat /proc/interrupts   # inspect your /proc/interrupts file for multiple devices having the same interrupt

$ lspci # lists all PCI buses and devices connected to them
$ lsusb # lists all USB buses and any connected USB devices
$ lshal # lists all devices the hardware abstraction layer (HAL) knows about (should be most hardware on your system)
$ lshw  # lists hardware on your system, including maker, type, and where it is connected

$ uname -a  # system architecture
$ cat /etc/issue                   # Display distribution and version (on some distros)
$ lsb_release -a
$ grep ^VERSION /etc/SuSE-release  # To see which SuSE Linux version you are using
$ cat /etc/mandriva-release        # To see which Mandriva Linux version you are using
$ dmesg | head                     # full version info.
$ cat /proc/version                # full version info.
$ cat /etc/issue                   # display Linux distribution
$ pstree
$ lsof | grep TCP                  # list open files
$ lsof |grep ' root ' |awk '{print $NF}' |sort|uniq|wc -l # list number of open files for a user
$ lsof -i :22 # list all connections via port 22 (i.e., ssh)
$ strace 'command'                 # trace system calls and signals for a given command

$ getconf     # print system configuration variables
$ getconfig   # get configuration information for the Xorg server
$ systool     # view system device information by bus, class, and topology
$ dmidecode   # DMI table decoder
$ dmidecode -s system-product-name
$ dmidecode -s system-manufacturer
$ biosdecode  # BIOS information decoder
$ bind -P     # print keyboard bindings

$ acpi -t     # Check current battery charge system temperature (package might not be installed by default)
$ finger -l   # Display information about all system users

$ cat /proc/scsi/scsi
  WDC WD2000JD-22H Rev: 08.0
  SATA-I, 200 GB, 150 MB/s, 8 MB Cache, 7200 RPM

Managing modules / devices / libraries / objects / etc

$ lspci
$ lsmod
$ depmod
$ modprobe  # tail /var/log/messages (to check success / failure)
$ modprobe -l |more   # list all the modules available for your kernel.
$ rmmod
$ ldd /path/to/library/file        # print shared library dependencies
$ nm /path/to/object/file          # list symbols from object files
$ nm [-s|--print-armap] /path/to/object/file # list index generated from a ranlib

$ ldd /usr/bin/python   # print shared library dependencies
       linux-gate.so.1 =>  (0xffffe000)
       libpython2.5.so.1.0 => /usr/lib/libpython2.5.so.1.0 (0xb7e2e000)
       libpthread.so.0 => /lib/libpthread.so.0 (0xb7e16000)
       libdl.so.2 => /lib/libdl.so.2 (0xb7e12000)
       libutil.so.1 => /lib/libutil.so.1 (0xb7e0e000)
       libm.so.6 => /lib/libm.so.6 (0xb7de7000)
       libc.so.6 => /lib/libc.so.6 (0xb7cb9000)
       /lib/ld-linux.so.2 (0xb7f71000)

Default runlevel

It is a good idea to make the default runlevel for your machine "3" (i.e. full multiuser mode without X11). This will prevent your system from hanging if something is wrong with your X11 settings (the graphics).

To change the default runlevel, edit your /etc/inittab file and change the line that reads

id:5:initdefault:

to

id:3:initdefault:

Now, everytime you turn on your machine (or reboot it), you will be taken to a CLI. Login as a user (not root!) and enter the following:

$ startx

see also: wikipedia:init

Linux networking

see: SuSE wireless card configuration

$ hostname -i           # show current IP address
$ hostname -d           # show current domain name
$ domainname            # show full domain name
$ traceroute
$ mtr
$ bwm-ng
$ dig
$ cat /etc/hosts        # show host configuration
$ cat /etc/sysconfig/network   # show gateway configuration
$ cat /etc/resolv.conf  # show DNS configuration (aka "nameserver(s)"; one per line)
$ cat /etc/iftab    # show MAC address (and various network interfaces; only for some distros)
$ cat /proc/net/arp     # show MAC address (and various network interfaces)
$ arp                   # manipulate the system ARP cache

$ ip a show
$ /etc/init.d/network restart
$ route add 20.0.xxx.xxx gateway foo
$ /etc/rc.local
$ /etc/sysconfig/network-scripts
$ /sbin/ifconfig

$ ethtool -s eth0 speed 100 duplex full autoneg off  # force full-speed traffic
$ ethtool eth0   # to check that it worked
$ netstat -ivn   # for tuning
$ mii-tool --force=100baseTx-FD eth0   # obsolete way

$ netstat -nr
Kernel IP routing table
Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
20.0.xxx.xx     20.0.xx.xx      255.255.255.0   UGH       0 0          0 eth0

$ cat /proc/net/arp
IP address       HW type     Flags       HW address            Mask     Device
192.168.xxx.xxx  0x1         0x2         00:00:00:00:00:00     *        eth0
192.168.xxx.xxx  0x1         0x2         00:00:00:00:00:00     *        eth0

$ netstat -plant  # extremely useful for troublshooting

NFS

Check your /etc/exports for directories mountable by IP address. E.g.,

/mnt/disk/data 10.0.67.53(rw) 10.0.67.123(ro)

Then execute the following (as root):

/usr/sbin/exportfs -a

Add the following line (for static routes) to you /etc/sysconfig/network-scripts/route-eth0

10.0.34.54 via 10.0.67.43

You can also accomplish the above via the CLI:

ip route add 10.0.34.54 via 10.0.67.43 dev eth0

Use /etc/sysconfig/network for your default gateway. E.g.,

HOSTNAME=foo.bar.com
NETWORKING=yes
GATEWAY=10.0.54.123
GATEWAYDEV=eth0

Logging

If you are getting a bunch of

martian destination 0.0.0.0 from xxx.xxx.xxx.xxx, dev eth0

messages in your logs (check dmesg |grep martian), you can turn this off by editing your /etc/sysctl.conf and changing:

net.ipv4.conf.all.log_martians=1
~ TO ~
net.ipv4.conf.all.log_martians=0

Follow or watch the httpd live fullstatus/requests

watch -n1 "cat /proc/loadavg && free -m|grep / && service httpd fullstatus|egrep 'GET|POST|VHost|request'"

External resources

Force umount when the "device is busy"

$ fuser -km /mnt/hda1

Display (monitor / graphics card)

$ cat /etc/X11/xorg.conf
$ cat /etc/X11/xdm/Xservers  # lists commands used to start the local X-server
$ xdpyinfo | grep dimen      # for screen dimensions

If you are having trouble (in SuSE) getting your monitor to display anything (either from an initial boot or from adding a new monitor), try the following:

Reboot in Failsafe mode
Login as root
At the command prompt type: sax2 -m 0=vesa
Configure video settings and test them (it is important to test your settings first!)
Reboot in normal mode.

External links

Backing up the MBR

WARNING: This article or section describes techniques which can be dangerous for your computer's hardware or the data on it. I have tested everything I describe herein on my personal computers. However, absolutely no guarantee can be made that it will work for you. Proceed with caution!

It is easy to backup and restore the master boot record (MBR) in Linux. However, caution must be exorcised when performing any of the following commands.

to backup

$ dd if=/dev/xxx of=mbr.backup bs=512 count=1

to restore

$ dd if=mbr.backup of=/dev/xxx bs=512 count=1

where xxx is the device, which can be hda, sda, or any other.

Sound problems

Note, my sound card specs: Intel Corporation 82801EB/ER (ICH5/ICH5R) AC'97 Audio Controller

Un-mute PCM sound
Check the following:

$ lsmod | grep snd
$ cat /etc/modprobe.conf
$ vi /etc/modprobe.d/sound

(change "snd-intel8x0" to "snd_intel8x0")

$ ./sbin/lspci

(list sound specs / Multimedia audio controller)

$ modprobe snd-via82xx

Configuring a Firewire (IEEE1394) interface

Check that the file /etc/modules.conf (for 2.4 kernels) or /etc/modprobe.conf (for 2.6 kernels) contains the line:

alias ieee1394-controller ohci1394

Note: If you do not have a /etc/modprobe.conf file, there is a utility to create one. As root, type:

$ /sbin/generate-modprobe.conf > /etc/modprobe.conf

Caution: This will overwrite the previous file (if one existed), so it would be best to back it up first. Take extra caution that it has not altered your previous (working) video and/or sound driver configurations!

You might need to reboot with your Firwire plugged in (if hotplugging is not setup) and check that the card is recognised and the module loaded with the following command (as root):

$ lsmod | grep 1394
  ohci1394               32240  0
  ieee1394              286264  1 ohci1394

Repair corrupted .Xauthority file

$ mkxauth -u user -c

Adobe acroread "adobe expr: syntax error"

For some reason, after installing the latest version of Adobe Reader 7.0.9 for Linux (aka acroread; 2007-04-11) and running it, I get an infinite loop of "expr: syntax error".

After digging around Google a bit, I found a simple solution (not sure if this is the best one). /usr/bin/acroread is just a Bourne shell script text executable. Open this script and replace the following

$ echo $mfile| sed 's/libgtk-x11-([0-9]*).0.so.0.([0-9])00.([0-9]*)|(.*)/123/g'
# ~OR~
$ echo $mfile| sed 's/libgtk-x11-\([0-9]*\).0.so.0.\([0-9]\)00.\([0-9]*\)\|\(.*\)/\1\2\3/g'

with the following

$ echo $mfile| sed 's/libgtk-x11-([0-9]*).0.so.0.([0-9]*)00.([0-9]*)|(.*)/123/g'
# ~OR~
$ echo $mfile| sed 's/libgtk-x11-\([0-9]*\).0.so.0.\([0-9]*\)00.\([0-9]*\)\|\(.*\)/\1\2\3/g'

It is just the second 'match all digits' regex "*" symbol missing.

That should do it. Not sure why, how, or if this is the problem, but it seems to work just fine on my machine (Note: openSuSE didn't need this fix; Mandriva Linux 2007.0 did).

Bootsplash

I like to turn off my bootsplash (otherwise, I am always hitting the "Esc" key). This can be accomplished by setting the /etc/sysconfig/bootsplash file to:

SPLASH="no"  # disables bootup graphics

Further controls can be found in your /etc/bootsplash/themes/*/config/ directory. For an example, in SuSE it is located here:

/etc/bootsplash/themes/SuSE/config/
/etc/bootsplash/themes/SuSE/config/bootsplash-1440x900.cfg  # example cfg file

Hacked?

Check for failed logins in: /var/log/messages
Regularly monitor:
- ```
zcat /var/log/auth.log.*.gz | grep refused
```
- ```
grep -i failed /var/log/auth.log
```
- last (successful logins) / lastb (unsuccessful logins)
- w and/or who
- uptime
Verify that /etc/passwd has not changed.
Check fuser for ports.
Search for portscans in server report.
Check for weird processing hogging the CPU.
Install and use rkhunter
Use fail2ban, DenyHosts, etc.

Notes

pstree — display a tree of processes
lsmod — program to show the status of modules in the Linux Kernel
modprobe — program to add and remove modules from the Linux Kernel
netstat — Print network connections, routing tables, interface statistics, masquerade connections, and multicast memberships
lspci — list all PCI devices
more /usr/share/pci.ids — A list of all known PCI ID's (vendors, devices, classes, and subclasses). Maintained at The Linux PCI ID Repository, use the update-pciids utility to download the most recent version.

External links

This category currently contains no pages or media.

Difference between revisions of "Category:Linux troubleshooting"

Latest revision as of 19:11, 6 March 2015

Contents

Emergency reboot

System information

See also

Managing modules / devices / libraries / objects / etc

Default runlevel

Linux networking

NFS

Logging

External resources

Force umount when the "device is busy"

Display (monitor / graphics card)

External links

Backing up the MBR

Sound problems

Configuring a Firewire (IEEE1394) interface

Repair corrupted .Xauthority file

Adobe acroread "adobe expr: syntax error"

Bootsplash

Hacked?

See also

See also

Notes

External links

Navigation menu

Personal tools

Namespaces

Variants

Views

More

Search

Navigation

Tools