The Empirical Manifesto

Trying to undo a corporate corner I painted myself into fifteen years back...

I wrote a webpage:

cat index.html | wc --> 3,827 characters

Used wkhtmltopdf to convert it to PDF, then used Word to turn it back into a webpage:

cat index.html | wc --> 6,646,652 characters

Travel ready; new disk added to pool and full remote access to my systems:

When macOS can nslookup, but returns the following error when attempting ssh/ping:

ssh: Could not resolve hostname unraid: nodename nor servname provided, or not known

Simply run:

% sudo killall -HUP mDNSResponder

rmt:/var/lib/rmt/public # for x in `find . -name "*aarch64*" | sed -r 's/.{2}//'` ; do rm -Rf $x ; done
rmt:/var/lib/rmt/public # for x in `find . -name "*s390x*" | sed -r 's/.{2}//'` ; do rm -Rf $x ; done
rmt:/var/lib/rmt/public # for x in `find . -name "*ppc64le*" | sed -r 's/.{2}//'` ; do rm -Rf $x ; done

Then rmt-cli mirror just the ones you need for quick/dirty:

rmt: # for x in `rmt-cli repos list --all | grep 15 | grep SP5 | grep -v Not | grep -v Don | awk '{print $2}'` ; do rmt-cli mirror repository $x ; done

zypper ref | grep "gpg-pubkey-" | xargs rpm -e

ehowton@wdt:~> crontab -l
*/5 * * * * /home/ehowton/scripts/directory_listing_unixspook
*/5 * * * * /home/ehowton/scripts/directory_listing_wdt
*/5 * * * * /home/ehowton/scripts/rotate_lj_pic_banner.sh

ehowton@wdt:~/scripts> cat directory_listing_wdt
#!/bin/bash

rm /srv/www/vhosts/wilddamntexan/kids/index.html
echo "" > /srv/www/vhosts/wilddamntexan/kids/index.html
for x in `ls -last /srv/www/vhosts/wilddamntexan/kids/ | awk '{print $10;}'`;
do echo "$x
" >> /srv/www/vhosts/wilddamntexan/kids/index.html;
done

ehowton@wdt:~/scripts> cat directory_listing_unixspook
#!/bin/bash

rm /srv/www/vhosts/unixspook/drop/index.html
echo "" > /srv/www/vhosts/unixspook/drop/index.html
for x in `ls -last /srv/www/vhosts/unixspook/drop/ | awk '{print $10;}'`;
do echo "$x
" >> /srv/www/vhosts/unixspook/drop/index.html;
done

ehowton@wdt:~/scripts> cat rotate_lj_pic_banner.sh
#!/bin/bash
# Script which rotates my LiveJournal header image

cp /srv/www/vhosts/wilddamntexan/kids/kansas1.jpg /srv/www/vhosts/wilddamntexan/kids/spacer.jpg &
sleep 60
cp /srv/www/vhosts/wilddamntexan/kids/kansas2.jpg /srv/www/vhosts/wilddamntexan/kids/spacer.jpg &
sleep 60
cp /srv/www/vhosts/wilddamntexan/kids/kansas3.jpg /srv/www/vhosts/wilddamntexan/kids/spacer.jpg &
sleep 60
cp /srv/www/vhosts/wilddamntexan/kids/kansas4.jpg /srv/www/vhosts/wilddamntexan/kids/spacer.jpg &
sleep 60
cp /srv/www/vhosts/wilddamntexan/kids/kansas5.jpg /srv/www/vhosts/wilddamntexan/kids/spacer.jpg &

exit

for x in `cat hostnames`; do echo $(nslookup $x | grep -e Address | sed -n 2p | sed -r 's/.{9}//') $x ; done >> hostfile



for x in `cat hostnames`; do echo $(nslookup $x | grep -e Address | sed -n 2p | sed -r 's/.{9}//') $x $(echo $x | sed "s/\..*//") $(echo $x | sed "s/\..*//").homelab.local ; done >> hostfile

https://documentation.suse.com/suma/4.2/en/suse-manager/client-configuration/client-upgrades-major.html

* for worky-work add proxy=[proxy] for Kernel Options (may not need)

You need at least SLES 12 SP5 before you can proceed to SLES 15 SP3/4
Upgrading from SLES 12 SP5 is only supported via an offline upgrade.

# zypper packages --orphaned
# zypper packages remove [module module module] (for Unknown URL in Add-Ons)

Systems --> Autoinstallation --> [upload autoyast.xml] --> Next -->

Add to Distribution KERNEL OPTIONS:
autoupgrade=1
useonlinerepo
insecure=1

Add to Profile VARIABLES Text Box:
IPADDR=192.168.1.69
GATEWAY=192.168.1.1

On the cloned machine, change the hostname and IP addresses. Make sure /etc/hosts contains the changes you made and the correct host entries.

For distributions that support systemd: If your machines have the same machine ID, delete the file on each duplicated client and re-create it:

# rm /etc/machine-id
rm /var/lib/dbus/machine-id
dbus-uuidgen --ensure
systemd-machine-id-setup
For distributions that do not support systemd: Generate a machine ID from dbus:

---

You don't really need to do these unless errors:

rm /var/lib/dbus/machine-id
dbus-uuidgen --ensure
If your clients still have the same Salt client ID, delete the minion_id file on each client (FQDN will be used when it is regenerated on client restart):

rm /etc/salt/minion_id
Delete accepted keys from the onboarding page and the system profile from SUSE Manager, and restart the client with:

service salt-minion restart
Re-register the clients. Each client will now have a different /etc/machine-id and should be correctly displayed on the System Overview page.

Had no idea xeyes was considered a unix "power tool"

[root@princess ~]# dnf --enablerepo=powertools install xorg-x11-apps

Find the Hog

If your users keep using up all the space in your home directory, here is a way to apprehend the top offenders.

cd /home
du -ks *|sort -nr|pg

(Note: The "k" option may not be necessary in non-posix systems.)

This string will show you all the directory sizes in order, largest first. Now if you are going to do a little cleanup yourself, in one of those directories run:

ls -ls|sort -nr|pg

This will list files by size largest first. That way when you do your compress, move, or remove, you may actually reclaim a significant amount of space.

Was working in my openSUSE 15.3 desktop and rebooted. So far, so good. When the box came back up and presented with with the login GUI, attempting to login errored with, "Sorry, you have to log out a previous session first. Multiple sessions are not supported." I mean, I assumed multiple sessions weren't supported but as the box had *just* come back up, I was 100% certain there were no other sessions on the box.

To fix: Shift+Alt+F1, login as root, and issue:

loginctl terminate-user ehowton

Shift+Alt+F7 back to to the login screen and login was successful.

Weird.

When:
dnsmasq[454824]: dnsmasq: cannot open log /var/log/dnsmasq.log: Permission denied

Do:
ausearch -c 'dnsmasq' --raw | audit2allow -M my-dnsmasq
semodule -X 300 -i my-dnsmasq.pp
systemctl start dnsmasq

● dnsmasq.service - DNS caching server.
   Loaded: loaded (/usr/lib/systemd/system/dnsmasq.service; enabled; vendor preset: disabled)
   Active: failed (Result: exit-code) since Wed 2022-01-19 10:06:14 CST; 5min ago
  Process: 454824 ExecStart=/usr/sbin/dnsmasq -k (code=exited, status=3)
 Main PID: 454824 (code=exited, status=3)

Jan 19 10:06:14 bkpsrv systemd[1]: Started DNS caching server..
Jan 19 10:06:14 bkpsrv dnsmasq[454824]: dnsmasq: cannot open log /var/log/dnsmasq.log: Permission denied
Jan 19 10:06:14 bkpsrv dnsmasq[454824]: cannot open log /var/log/dnsmasq.log: Permission denied
Jan 19 10:06:14 bkpsrv dnsmasq[454824]: FAILED to start up
Jan 19 10:06:14 bkpsrv systemd[1]: dnsmasq.service: Main process exited, code=exited, status=3/NOTIMPLEMENTED
Jan 19 10:06:14 bkpsrv systemd[1]: dnsmasq.service: Failed with result 'exit-code'.
[root@bkpsrv log]# systemctl start dnsmasq
[root@bkpsrv log]# systemctl status dnsmasq
● dnsmasq.service - DNS caching server.
   Loaded: loaded (/usr/lib/systemd/system/dnsmasq.service; enabled; vendor preset: disabled)
   Active: active (running) since Wed 2022-01-19 10:12:14 CST; 1s ago
 Main PID: 454964 (dnsmasq)
   Memory: 916.0K
   CGroup: /system.slice/dnsmasq.service
           └─454964 /usr/sbin/dnsmasq -k

Jan 19 10:12:14 bkpsrv systemd[1]: Started DNS caching server..

Matt's vm firmware upgrade:

grab file systems to that will be upgraded.
# df -h | grep vx | awk '{print $6}' | grep -v vx > list

grab vxdg that will upgrade
# vxdg list |grep -v NAME |awk '{print $1}' > list2

check volume group version
# for i in $(cat list2); do vxdg -q list $i |grep version; done

upgrade volume groups
# for i in $(cat list2); do vxdg upgrade $i; done

Verify VG upgrade successful
# for i in $(cat list2); do vxdg -q list $i |grep version; done

Now upgrade file system versions, check version level for each.
# df -h |grep vx| awk '{print $1}' |grep -v tmpfs > vxfs.out
# for i in $(cat vxfs.out); do /opt/VRTS/bin/fstyp -v $i |grep version; done

Step upgrade each file system, if current version is 9 start with 10 and upgrade through 15.
# for i in $(cat list); do /opt/VRTS/bin/vxupgrade -n 10 $i; done
# for i in $(cat list); do /opt/VRTS/bin/vxupgrade -n 11 $i; done
# for i in $(cat list); do /opt/VRTS/bin/vxupgrade -n 12 $i; done
# for i in $(cat list); do /opt/VRTS/bin/vxupgrade -n 13 $i; done
# for i in $(cat list); do /opt/VRTS/bin/vxupgrade -n 14 $i; done
# for i in $(cat list); do /opt/VRTS/bin/vxupgrade -n 15 $i; done

Finally verify all vxfs are version 15 now
# for i in $(cat vxfs.out); do /opt/VRTS/bin/fstyp -v $i |grep version; done

IF YOU GET A SIMILAR ERROR:
UX:vxfs vxupgrade: ERROR: V-3-27659: /sapmnt/SID is not a mountpoint
and the raw device cannot be deduced.
use [-r] to specify a raw device

Find the *actual* path via /etc/fstab & /etc/auto.direct and upgrade that path separately:

# /opt/VRTS/bin/vxupgrade -n 10 /export/sapmnt/SID
# /opt/VRTS/bin/vxupgrade -n 11 /export/sapmnt/SID
. . .
. . .
. . .
# /opt/VRTS/bin/vxupgrade -n 15 /export/sapmnt/SID

Originally posted by fgrehk @ Virten.net

Connect to the ESXi host with SSH
###########################################################################################################################
Stop the USB arbitrator service. This service is used to passthrough USB device from an ESX/ESXi host to a virtual machine.
(When disabling it, you can no longer passthrough USB devices to VMs)
-------------------------------------------
# /etc/init.d/usbarbitrator stop
-------------------------------------------
###########################################################################################################################
(optional) Use this command to permanently disable the USB arbitrator service after reboot.
-------------------------------------------
# chkconfig usbarbitrator off
-------------------------------------------
###########################################################################################################################
Plug in the USB device to your ESXi host. While connecting the USB device you can either watch /var/log/vmkernel.log
to identify the device name or identify it within /dev/disks. Devices are either named mpx.vmhbaXX or with an naa.X number.
-------------------------------------------
# ls /dev/disks/
-------------------------------------------
###########################################################################################################################
Write a GPT label to the device (Assuming that the Device ID is naa.5000000000000001)
--------------------------------------------------------
# partedUtil mklabel /dev/disks/naa.5000000000000001 gpt
--------------------------------------------------------
###########################################################################################################################
To create a partition you need to know the start sector, end sector, which depends on the device size and the GUID.
The start sector is always 2048
The GUID for VMFS is AA31E02A400F11DB9590000C2911D1B8
The end sector can be calculated with the following formula (Use the numbers from getptbl):
-----------------------------------------------------
# partedUtil getptbl /dev/disks/naa.5000000000000001
-----------------------------------------------------
gpt
15566 255 63 250069680
15566 * 255 * 63 - 1 = 250067789 (Your end sector)
###########################################################################################################################
You can also calculate the end sector with the following command:
------------------------------------------------------------------------------------------------------------------------
# eval expr $(partedUtil getptbl /dev/disks/naa.5000000000000001 | tail -1 | awk '{print $1 " \\* " $2 " \\* " $3}') - 1
------------------------------------------------------------------------------------------------------------------------
250067789 => Your end sector
###########################################################################################################################
Create the VMFS partition (Replace with your end sector)
--------------------------------------------------------------------------------------------------------------
# partedUtil setptbl /dev/disks/naa.5000000000000001 gpt "1 2048 250067789 AA31E02A400F11DB9590000C2911D1B8 0"
--------------------------------------------------------------------------------------------------------------
###########################################################################################################################
Format the partition with VMFS5 or VMFS6
VMFS v5
-------------------------------------------------------------------------
# vmkfstools -C vmfs5 -S USB-Datastore /dev/disks/naa.5000000000000001:1
-------------------------------------------------------------------------
VMFS v6
-------------------------------------------------------------------------
# vmkfstools -C vmfs6 -S USB-Datastore /dev/disks/naa.5000000000000001:1

NOW DO THIS:

esxcli system settings advanced set -o /Disk/AllowUsbClaimedAsSSD -I 1
esxcli storage nmp satp rule add -s VMW_SATP_LOCAL --device=mpx.vmhba32:C0:T0:L0 --option=enable_ssd
esxcli storage core claiming unclaim --type device --device=mpx.vmhba32:C0:T0:L0
esxcli storage core claimrule load
esxcli storage core claimrule run

REMEMBER TO:

In the vSphere Client, navigate to the ESXi host.
Click the Configure tab.
Under System, click Advanced System Settings.
In the Advanced System Settings table, select the Misc.APDTimeout parameter and click the Edit icon.
I've been using 600 as the value.

When, in VMware Fusion, the left-mouse click doesn't work, do this:

Shutdown all VMs.
Quit Fusion.

# vi /path/to/virtual_machine/vm.vmx

Add:

mouse.vusb.startConnected = "FALSE"

Fix you right up.

So this was a lot easier than trying to figure out how to get an XDMCP login propmt on the new flavors of linux login screen:

On the SGI box Toolchest --> System --> Utilities --> Enable Remote Display

Then on the linux host, in a console window (choose one):

yum install xnest
zypper install xnest
apt install xnest

Finally run this command:


Xnest -query host_to_connect -geometry 1280x1024 :1

To set a hostname on a server without an Enterprise iDRAC:

ssh user@IP

admin1> racadm config -g ifcRacManagedNodeOs -o ifcRacMnOsHostname hostname

for x in `cat hostnames`;
do echo $(nslookup $x | grep -e Address | sed -n 2p | sed -r 's/.{9}//') $x >> hostfile;
done