Skip to content

DAOS-18487 object: control EC rebuild resource consumption #17441

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
gnailzenh wants to merge 2 commits into
base: release/2.6
Choose a base branch
from
Open

DAOS-18487 object: control EC rebuild resource consumption #17441

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b280331
DAOS-18487 object: control EC rebuild resource consumption
gnailzenh Jan 24, 2026
987dc5c
DAOS-18487 object: degraded buffer size only impact resource control
gnailzenh Jan 26, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 18 additions & 2 deletions src/object/srv_obj_migrate.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1966,6 +1966,7 @@ migrate_one_ult(void *arg)
struct migrate_one *mrone = arg;
struct migrate_pool_tls *tls;
daos_size_t data_size;
daos_size_t degraded_size = 0;
int rc = 0;

while (daos_fail_check(DAOS_REBUILD_TGT_REBUILD_HANG))
Expand All @@ -1978,6 +1979,21 @@ migrate_one_ult(void *arg)
}

data_size = daos_iods_len(mrone->mo_iods, mrone->mo_iod_num);
if (daos_oclass_is_ec(&mrone->mo_oca)) {
/* NB: this is a workaround for EC object:
* The fetch buffer is taken from a pre-registered (R)DMA buffer;
* however, a degraded EC read will allocate and register an extra
* buffer to recover data.
*
* Currently, the resource manager cannot control this extra allocation,
* which can lead to increased memory consumption.
*
* While this workaround does not prevent dynamic buffer allocation and
* registration, it does provide relatively precise control over the
* resources consumed by degraded EC reads.
*/
degraded_size = data_size * MIN(8, obj_ec_data_tgt_nr(&mrone->mo_oca));
}
data_size += daos_iods_len(mrone->mo_iods_from_parity,
mrone->mo_iods_num_from_parity);

Expand All @@ -1986,13 +2002,13 @@ migrate_one_ult(void *arg)

D_ASSERT(data_size != (daos_size_t)-1);

rc = migrate_res_hold(tls, MIGR_DATA, data_size, NULL);
rc = migrate_res_hold(tls, MIGR_DATA, data_size + degraded_size, NULL);
if (rc)
D_GOTO(out, rc);

rc = migrate_dkey(tls, mrone, data_size);

migrate_res_release(tls, MIGR_DATA, data_size);
migrate_res_release(tls, MIGR_DATA, data_size + degraded_size);

D_DEBUG(DB_REBUILD, DF_UOID" layout %u migrate dkey "DF_KEY" inflight_size "DF_U64": "
DF_RC"\n", DP_UOID(mrone->mo_oid), mrone->mo_oid.id_layout_ver,
Expand Down
Loading