# Decomissioning hosts * Get rid of any services running on the host. * Optional, shut down the machine, then wait a month * Zero disks badblocks -v -s -p 5 -w -f DEVICE_GOES_HERE * Revoke puppet certificate * on handel: puppet node clean $HOST * reinit samhain on handel * Remove from LDAP * including references such as subgroup: foo@hostname or supplementarygid: foo@hostname * rerun puppet on backuphost / the backup host * reinit samhain on backuphost * Remove from Nagios * Remove any references from puppet configuration * Remove references to the host from DNS files, including reverse DNS * Get rid of the HW (tell the hoster to get rid of it or other as appropriate) * Update password file information for the $HOST * Remove MQ queues related to machine (see [[doc/mq]] for details) for queue in $(rabbitmqadmin -N rainier list queues name | \ grep unger.debian.org | \ awk '{print $2}') ; do rabbitmqadmin -N rainier -V dsa delete queue name=$queue done * Remove disks from the MSA * run multipath -ll on the KVM host and find all disks on that multipath ``` finzi-srv (3600c0ff00027786c874edd5701000000) dm-52 HP,MSA 2040 SAN size=75G features='2 queue_if_no_path retain_attached_hw_handler' hwhandler='0' wp=rw |-+- policy='service-time 0' prio=50 status=active | |- 1:0:2:4 sdeu 129:96 active ready running | |- 1:0:3:4 sdhu 134:64 active ready running | |- 2:0:0:4 sdaa 65:160 active ready running | `- 2:0:2:4 sdfz 131:80 active ready running `-+- policy='service-time 0' prio=10 status=enabled |- 1:0:0:4 sdf 8:80 active ready running |- 1:0:1:4 sdbu 68:128 active ready running |- 2:0:1:4 sdcz 70:112 active ready running `- 2:0:3:4 sdja 8:320 active ready running ``` * for each disk in those multipath run echo 1 > /sys/block/sdX/device/delete * once done, remove that multpath-device from multipath multipath -f finzi-srv * remove the paths in the MSA Three months after decomissioning: * Remove backups