ZDLRA ERROR: Filesystem label ESP is found more than once.

Recently, while patching a ZDLRA X7-2 full stack to version 21.1.202405, I encountered an error during the dbnode precheck. This blog post details the issue and its resolution.

The Error

During the precheck, the following error was encountered:


2024-09-17 15:13:11 +0200 :ERROR : dbnodeupdate.sh precheck failed on one or more nodes SUMMARY OF WARNINGS AND ERRORS FOR radb02-m: radb02-m: ERROR Filesystem label ESP is found more than once. Check multipath configuration and Contact Oracle Support

Investigation Initial Attempt

Upon investigation, I found a similar issue documented in the Oracle Support note: “Exadata Patchmgr Report ERROR: Filesystem Label ESP Is Found More Than Once (Doc ID 3034775.1)”.Checking the system, I confirmed the presence of duplicate ESP entries:

root@spra51db02-m ~]# lsblk -o name,mountpoint,label,size,uuid | grep ESP
├─sda2 ESP 256M C944-F751
├─3600062b20131c2002296d6fc12398fb5p2 /boot/efi ESP 256M C944-F751

Initial Attempt

Following the recommendation in the support document, I attempted to patch with Exadata release 24.1.1.0.0.240605. However, this did not resolve the issue, and the same error persisted.

Resolution

After contacting Oracle Support, we implemented the following solution:

  1. Remove the duplicate entry from the multipath configuration file /etc/multipath/wwids.
  2. Recreate the initramfs.

The detailed procedure for this resolution is as follows:

> Get WWID of /dev/sda
# multipath -v4 2>&1 | grep sda
> Remove the reported WWID from this file /etc/multipath/wwids
> Backup existing initramfs
# cd /boot
# cp -p initramfs-$(uname -r).img initramfs-$(uname -r).img.bkp
> Check the list of multipath files/dirs
# lsinitrd /boot/initramfs-$(uname -r).img | grep -i multipath
> Recreate initramfs without multipath config files
# depmod -a
# dracut -f –omit “multipath” -v
> Check the list of multipath files/dirs
# lsinitrd /boot/initramfs-$(uname -r).img | grep -i multipath
> Reboot node and verify
# lsblk
# blkid | grep ESP

Conclusion

This issue highlights the importance of careful system configuration and the potential complexities in patching engineered systems like ZDLRA. Always ensure your system’s configuration is clean and consistent before attempting major updates.If you encounter similar issues, don’t hesitate to consult Oracle Support for guidance tailored to your specific environment.Remember to test thoroughly after applying any changes and before proceeding with the patching process.

Published by dbaliw

Highly experienced Oracle Database Administrator and Exadata Specialist with over 15 years of expertise in managing complex database environments. Skilled in cloud technologies, DevOps practices, and automation. Certified Oracle Cloud Infrastructure Architect and Oracle Certified Master with a strong background in performance tuning, high availability solutions, and database migrations.

Leave a comment