centos filesystem升級注意事項

CentOS Filesystem Upgrade Precautions

1. Backup Critical Data Before Any Operation

Backing up is the most critical step to prevent data loss during filesystem upgrades or conversions (e.g., ext3 to ext4). Use tools like rsync, tar, or cloud storage to back up user data, configuration files (/etc, /var), and application data. Verify backup integrity before proceeding.

2. Check Version Compatibility

Ensure the new filesystem version is compatible with your CentOS version and existing software stack. For example, upgrading from ext3 to ext4 requires CentOS 7+ (ext4 support was added in CentOS 7), and newer ext4 features may not work on older kernels. For HDFS upgrades, verify compatibility with Hadoop components (e.g., Hadoop version, YARN) to avoid metadata corruption or service failures.

3. Perform Pre-Upgrade Checks

Use tools like preupgrade-assistant (for system upgrades) or e2fsck (for filesystem checks) to identify potential issues. Run e2fsck -f /dev/sdXn to check and repair filesystem errors before converting (replace /dev/sdXn with your actual partition). For system upgrades, preupgrade-assistant-cli helps detect unsupported packages or configurations.

4. Choose the Right Upgrade Method

• System Upgrade: Use dnf (CentOS 8+) or yum (CentOS 7) to upgrade the entire system, including the filesystem. Example: sudo dnf upgrade --releasever=8 for CentOS 7→8.
• Filesystem Conversion: For ext3→ext4, use tune2fs and fsck without reformatting:

  sudo tune2fs -O extents,uninit_bg /dev/sdXn  # Enable ext4 features
  sudo fsck -f /dev/sdXn                      # Check filesystem
  sudo mount -o remount,rw /dev/sdXn          # Remount as read-write
  

  Update /etc/fstab with ext4 as the filesystem type.

5. Handle LVM and Disk Space Carefully

If using LVM, extend logical volumes (lvextend) before resizing the filesystem (resize2fs). For non-LVM setups, ensure sufficient unallocated disk space (at least 10-20% of the filesystem size) to accommodate new features or growth. Avoid overwriting existing partitions during conversion.

6. Update /etc/fstab Correctly

Modify /etc/fstab to reflect the new filesystem type and optimal mount options. For ext4, common options include defaults,noatime,nodiratime (disable access time updates for better performance). Example entry:

/dev/sdXn  /mnt/data  ext4  defaults,noatime,nodiratime  0  2

Double-check UUIDs (use blkid to verify) if using UUID-based mounting.

7. Monitor and Validate Post-Upgrade

After upgrading, reboot the system and verify:

• Filesystem type: df -Th | grep sdXn (should show ext4).
• Mount status: mount | grep sdXn (should be mounted with correct options).
• Data integrity: Check logs (/var/log/messages, journalctl -xe) for errors and test application functionality.

8. Prepare a Rollback Plan

For HDFS upgrades, use the -rollback option to revert to the old version if issues arise. For system upgrades, keep a backup of the original repository files (/etc/yum.repos.d/*.repo) and a list of installed packages (rpm -qa > package-list.txt) to restore the previous state if needed.

9. Address Common Issues Proactively

• Cannot Unmount Filesystem: Use lsof | grep /mount-point or fuser -m /mount-point to find and kill processes using the filesystem before unmounting.
• e2fsck Errors: Run e2fsck -p /dev/sdXn (automatic repair) or e2fsck -y /dev/sdXn (interactive repair) to fix corruption.
• Performance Degradation: Optimize /etc/fstab with options like data=writeback (for better write performance) or noatime (for reduced disk I/O).