Re: [pbs-devel] [PATCH proxmox-backup v2] api: zfs: create zpool with relatime=on

[Thomas Lamprecht <t.lamprecht@proxmox.com>]

On 03.11.21 08:36, Dominik Csapak wrote:
> some operations (e.g. garbage collection/restore/etc.) are very read
> intensive on the chunks, and having atime=on and relatime=off (zfs default)
> makes those write intensive operations too. Additionally, 'ext4' defaults to
> relatime, so also change the default for api-created zpools.
> 
> Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
> ---
> changes from v1:
> * do not disable atime, it must be on for relatime to work, quote
> from 'man zfsprops' for 'relatime':
> 
> Controls the manner in which the access time is updated when atime=on is set.
> 
>  src/api2/node/disks/zfs.rs | 12 +++++++-----
>  1 file changed, 7 insertions(+), 5 deletions(-)
> 
> diff --git a/src/api2/node/disks/zfs.rs b/src/api2/node/disks/zfs.rs
> index 9c3a0958..56dedab5 100644
> --- a/src/api2/node/disks/zfs.rs
> +++ b/src/api2/node/disks/zfs.rs
> @@ -273,13 +273,15 @@ pub fn create_zpool(
>                  proxmox_systemd::enable_unit(&import_unit)?;
>              }
>  
> +            let mut command = std::process::Command::new("zfs");
> +            command.arg("set");
>              if let Some(compression) = compression {
> -                let mut command = std::process::Command::new("zfs");
> -                command.args(&["set", &format!("compression={}", compression), &name]);
> -                task_log!(worker, "# {:?}", command);
> -                let output = pbs_tools::run_command(command, None)?;
> -                task_log!(worker, "{}", output);
> +                command.arg(&format!("compression={}", compression));
>              }
> +            command.args(&["relatime=on", &name]);
> +            task_log!(worker, "# {:?}", command);
> +            let output = pbs_tools::run_command(command, None)?;
> +            task_log!(worker, "{}", output);
>  
>              if add_datastore {
>                  let lock = pbs_config::datastore::lock_config()?;

looks good to me but would need a rebase on current master branch.