Hello, friend!
The post will share with you the dusk failure during cluster preinstallation.
When you run Preinstall during cluster installation, the system displays a message indicating that Disk format failed.
The PreInstall fails due to disk formatting failure.
Cause Analysis
[Fri Jun 5 10:15:30 CST 2021] [INFO] [get_blank_disk:1733] dev(sda) exists in mount.^M [Fri Jun 5 10:15:30 CST 2021] [INFO] [get_blank_disk:1733] dev(sdb) exists in mount.^M [Fri Jun 5 10:15:30 CST 2021] [INFO] [match_dev_size:1892] dev(sdc) size(99G) can match min required size(1G).^M [Fri Jun 5 10:15:30 CST 2021] [INFO] [disk_line_ready:2016] ready line conf(datanode1.conf) match dev(sdc) without condition.^M [Fri Jun 5 10:15:30 CST 2021] [ERROR] [disk_line_ready:2053] ready line conf(datanode2.conf) can not match dev.^M [Fri Jun 5 10:15:30 CST 2021] [ERROR] [disk_ready:2093] disk ready line(datanode2.conf y y n ) invalid.^M [Fri Jun 5 10:15:30 CST 2021] [ERROR] [main:3300] disk ready failed.^M [Fri Jun 5 10:15:30 CST 2021] [INFO] [rollback_mount:2943] rollback no mount any yet.^M [Fri Jun 5 10:15:30 CST 2021] [ERROR] [none:3527] execute autopart.sh failed.^M
Analyze the PreInstall log (/var/log/fi-preinstall.log)
The free disk of the server is insufficient. As a result, datanode2 fails to be mounted.
Solution
If the planned disk has been mounted to the system, detach the disk and run the preinstall command again.
If the disk space of the server is insufficient, add disks or adjust the cluster plan to reduce the number of disks attached to the server.
That's all, thanks!
