c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2025-02-04 14:24:22 +03:00
Code Issues Packages Projects Releases Wiki Activity

ddt: tuneable to override copies= on dedup metadata objects

Browse Source 
All objects stored in the MOS get copies=3. For a large dedup table,
this requires significant extra IO and disk space, when its not really
necessary - the dedup table itself isn't needed to read or write data,
only to keep data usage down. Losing the dedup table does not render the
pool unusable, it just messes up the accounting somewhat.

This adds a dmu_ddt_copies tuneable. When set to 0, the existing
behaviour is used. When set higher, dedup table blocks (ZAP and log)
will have this many copies rather than the usual 3, while indirect
blocks will have one more again.

This is a tuneable for now mostly for testing. Losing a dedup table can
cause blocks to be leaked, and we currently have no facilities to repair
that.

Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Co-authored-by: Allan Jude <allan@klarasystems.com>
Signed-off-by: Rob Norris <rob.norris@klarasystems.com>
Sponsored-by: Klara, Inc.
Sponsored-by: iXsystems, Inc.
Closes #15895
This commit is contained in:
Rob Norris 2023-10-06 17:06:34 +11:00 committed by Brian Behlendorf
parent 592f38900d
commit cbb9ef0a4c
1 changed files with 32 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
module/zfs/dmu.c
View File

@ -95,6 +95,12 @@ uint_t dmu_prefetch_max = 8 * 1024 * 1024;
uint_t dmu_prefetch_max = 8 * SPA_MAXBLOCKSIZE;
#endif
/*
* Override copies= for dedup state objects. 0 means the traditional behaviour
* (ie the default for the containing objset ie 3 for the MOS).
*/
uint_t dmu_ddt_copies = 0;
const dmu_object_type_info_t dmu_ot[DMU_OT_NUMTYPES] = {
{DMU_BSWAP_UINT8, TRUE, FALSE, FALSE, "unallocated" },
{DMU_BSWAP_ZAP, TRUE, TRUE, FALSE, "object directory" },
@ -2272,6 +2278,28 @@ dmu_write_policy(objset_t *os, dnode_t *dn, int level, int wp, zio_prop_t *zp)
case ZFS_REDUNDANT_METADATA_NONE:
break;
}
if (dmu_ddt_copies > 0) {
/*
* If this tuneable is set, and this is a write for a
* dedup entry store (zap or log), then we treat it
* something like ZFS_REDUNDANT_METADATA_MOST on a
* regular dataset: this many copies, and one more for
* "higher" indirect blocks. This specific exception is
* necessary because dedup objects are stored in the
* MOS, which always has the highest possible copies.
*/
dmu_object_type_t stype =
dn ? dn->dn_storage_type : DMU_OT_NONE;
if (stype == DMU_OT_NONE)
stype = type;
if (stype == DMU_OT_DDT_ZAP) {
copies = dmu_ddt_copies;
if (level >=
zfs_redundant_metadata_most_ditto_level)
copies++;
}
}
} else if (wp & WP_NOFILL) {
ASSERT(level == 0);
@ -2824,3 +2852,7 @@ ZFS_MODULE_PARAM(zfs, zfs_, dmu_offset_next_sync, INT, ZMOD_RW,
/* CSTYLED */
ZFS_MODULE_PARAM(zfs, , dmu_prefetch_max, UINT, ZMOD_RW,
"Limit one prefetch call to this size");
/* CSTYLED */
ZFS_MODULE_PARAM(zfs, , dmu_ddt_copies, UINT, ZMOD_RW,
"Override copies= for dedup objects");