zfs: don't specify zfs arc cache

Turns out, zfs is smart!
ZFS already has sane defaults, no sense in limiting the size of the cache.

modules/zfs.nix

@@ -8,25 +8,8 @@
   boot.zfs.package = pkgs.zfs;
   boot.initrd.kernelModules = [ "zfs" ];
 
-  boot.kernelParams =
-    let
-      arc_gb = 32;
-      arc_mb = arc_gb * 1000;
-      arc_kb = arc_mb * 1000;
-      arc_b = arc_kb * 1000;
-
-      dirty_gb = 8; # Default value is 4GB, helps smooth writes
-      dirty_mb = dirty_gb * 1000;
-      dirty_kb = dirty_mb * 1000;
-      dirty_b = dirty_kb * 1000;
-    in
-    [
-      "zfs.zfs_arc_max=${builtins.toString arc_b}"
+  boot.kernelParams = [
     "zfs.zfs_txg_timeout=120" # longer TXG open time = larger sequential writes
-      "zfs.zfs_dirty_data_max=${builtins.toString dirty_b}"
-      "zfs.zfs_delay_min_dirty_percent=80" # delay write throttling until 80% dirty (default 60%)
-      "zfs.zfs_vdev_async_write_max_active=30" # more concurrent async writes to vdevs (default 10)
-      "zfs.zfs_vdev_async_read_max_active=10" # more concurrent async reads for random I/O (default 3)
   ];
 
   boot.supportedFilesystems = [ "zfs" ];