How to use the command 'zdb' (with examples)

The ‘zdb’ command is a ZFS (Z File System) debugger that provides detailed configuration and statistics about ZFS zpools. This command is useful for troubleshooting and gathering information about ZFS pools.

Use case 1: Show detailed configuration of all mounted ZFS zpools

Code:

zdb

Motivation: By using the command ‘zdb’ without any arguments, we can retrieve detailed configuration information for all mounted ZFS zpools. This can be helpful for system administrators to understand the current state of the zpools and analyze any issues or performance concerns.

Explanation: The ‘zdb’ command without any arguments fetches and displays the configuration information and statistics of all mounted ZFS zpools. It provides information such as pool name, pool address, pool guid, and various other attributes related to the zpool.

Example output:

dnode flags = 0xa000000000000000
compiled hostid = 0x01983383
        ipfs ino object: jroot
        filetype: regular file
        ... (additional output)

Use case 2: Show detailed configuration for a specific ZFS pool

Code:

zdb -C poolname

Motivation: In certain cases, we might want to retrieve detailed configuration information for a specific ZFS pool instead of all mounted zpools. This can be useful when troubleshooting issues or analyzing performance concerns specific to a particular zpool.

Explanation: The ‘-C’ option followed by the pool name allows us to retrieve detailed configuration information for a specific ZFS pool. This narrows down the output to only the specified zpool, providing detailed information such as pool name, pool address, pool guid, and various other attributes related to the zpool.

Example output:

Uberblock[0] = v22258
        magic = 00000005
        version = 5000
        txg_checkpoint = 6791023
        txg_syncing = 6791023
        txg = 6791135
        guid_sum = 123456789
        timestamp = 1564319901

Use case 3: Show statistics about number, size, and deduplication of blocks

Code:

zdb -b poolname

Motivation: When dealing with ZFS zpools, it can be helpful to gather statistics about the number, size, and deduplication of blocks to understand the efficiency and capacity utilization of the zpool. This information can assist in identifying any potential data deduplication benefits or space optimization opportunities.

Explanation: The ‘-b’ option followed by the pool name allows us to retrieve statistics related to the number, size, and deduplication of blocks within a ZFS pool. This includes details such as total number of blocks, amount of space used by deduplicated blocks, and other statistics related to block allocation and utilization.

Example output:

Metaslabs Summary report:
    Implemented metaslabs boost:  TRUE
    Linear metaslab groups:          1
    ...
    Allocatable metaslabs:        928
    Share allocatable metaslabs:     928
    Average predefined metaslab size:   5722496
    Average freelist size:         24576

Conclusion:

The ‘zdb’ command is a powerful tool for troubleshooting and gathering information about ZFS zpools. By using different options, we can retrieve detailed configuration information for all mounted zpools, a specific zpool, or statistics about block usage. These insights help system administrators understand the state of their zpools and make informed decisions related to performance optimization and troubleshooting.