Skip to content

DAOS-18487: down2up fix #17570

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

Draft
gnailzenh wants to merge 1 commit into
base: master
Choose a base branch
from
+30 −13
Draft

DAOS-18487: down2up fix #17570

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
9 changes: 8 additions & 1 deletion src/placement/jump_map.c
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
/**
*
* (C) Copyright 2016-2024 Intel Corporation.
* (C) Copyright 2025 Hewlett Packard Enterprise Development LP
* (C) Copyright 2025-2026 Hewlett Packard Enterprise Development LP
*
* SPDX-License-Identifier: BSD-2-Clause-Patent
*/
Expand Down Expand Up @@ -734,6 +734,13 @@ get_object_layout(struct pl_jump_map *jmap, uint32_t layout_ver, struct pl_obj_l
setbit(dom_cur_grp_real, domain - root);
if (pool_target_down(target))
layout->ol_shards[k].po_rebuilding = 1;

if (pool_target_is_down2up(target)) {
if (gen_mode == PRE_REBUILD)
layout->ol_shards[k].po_rebuilding = 1;
else
layout->ol_shards[k].po_reintegrating = 1;
}
}

if (is_extending != NULL && pool_target_is_up_or_drain(target))
Expand Down
24 changes: 14 additions & 10 deletions src/placement/pl_map_common.c
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
/**
* (C) Copyright 2016-2024 Intel Corporation.
* (C) Copyright 2026 Hewlett Packard Enterprise Development LP
*
* SPDX-License-Identifier: BSD-2-Clause-Patent
*/
Expand Down Expand Up @@ -255,13 +256,7 @@ is_comp_avaible(struct pool_component *comp, uint32_t allow_version,
status = PO_COMP_ST_UPIN;
} else if (status == PO_COMP_ST_UP) {
if (comp->co_flags & PO_COMPF_DOWN2UP) {
/* PO_COMP_ST_UP status with PO_COMPF_DOWN2UP flag
* is the case of delay_rebuild exclude+reint.
* Cannot mark it as UPIN to avoid it be used for
* rebuild enumerate/fetch, as the data will be
* discarded in reintegrate.
*/
/* status = PO_COMP_ST_UPIN; */
status = PO_COMP_ST_UPIN;
} else {
if (comp->co_fseq <= 1)
status = PO_COMP_ST_NEW;
Expand Down Expand Up @@ -394,9 +389,14 @@ determine_valid_spares(struct pool_target *spare_tgt, struct daos_obj_md *md,
if (spare_avail) {
/* The selected spare target is up and ready */
l_shard->po_target = spare_tgt->ta_comp.co_id;
l_shard->po_fseq = f_shard->fs_fseq;
l_shard->po_rank = spare_tgt->ta_comp.co_rank;
l_shard->po_index = spare_tgt->ta_comp.co_index;
l_shard->po_fseq = f_shard->fs_fseq;
l_shard->po_rank = spare_tgt->ta_comp.co_rank;
l_shard->po_index = spare_tgt->ta_comp.co_index;

if (pool_target_is_down2up(spare_tgt))
f_shard->fs_down2up = 1;
else
f_shard->fs_down2up = 0;

/*
* Mark the shard as 'rebuilding' so that read will skip this shard.
Expand All @@ -406,6 +406,10 @@ determine_valid_spares(struct pool_target *spare_tgt, struct daos_obj_md *md,
f_shard->fs_status == PO_COMP_ST_DRAIN ||
f_shard->fs_down2up || pool_target_down(spare_tgt))
l_shard->po_rebuilding = 1;

if (f_shard->fs_down2up && gen_mode != PRE_REBUILD)
l_shard->po_reintegrating = 1;

} else {
l_shard->po_shard = -1;
l_shard->po_target = -1;
Expand Down
10 changes: 8 additions & 2 deletions src/rebuild/srv.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1998,16 +1998,18 @@ rebuild_task_ult(void *arg)
uint32_t map_dist_ver = 0;
struct rebuild_global_pool_tracker *rgt = NULL;
d_rank_t myrank;
uint64_t cur_ts = 0;
// uint64_t cur_ts = 0;
uint32_t obj_reclaim_ver = 0;
int rc;

#if 0
cur_ts = daos_gettime_coarse();
D_ASSERT(task->dst_schedule_time != (uint64_t)-1);
if (cur_ts < task->dst_schedule_time) {
D_INFO("rebuild task sleep " DF_U64 " second\n", task->dst_schedule_time - cur_ts);
dss_sleep((task->dst_schedule_time - cur_ts) * 1000);
}
#endif

rc = ds_pool_lookup(task->dst_pool_uuid, &pool);
if (pool == NULL) {
Expand Down Expand Up @@ -2209,6 +2211,8 @@ rebuild_ults(void *arg)

while (!d_list_empty(&rebuild_gst.rg_queue_list) ||
!d_list_empty(&rebuild_gst.rg_running_list)) {
uint64_t now;

if (rebuild_gst.rg_abort) {
D_INFO("abort rebuilds\n");
break;
Expand All @@ -2222,13 +2226,15 @@ rebuild_ults(void *arg)
continue;
}

now = daos_gettime_coarse();
task = d_list_entry(rebuild_gst.rg_queue_list.next, struct rebuild_task, dst_list);
while (&rebuild_gst.rg_queue_list != &task->dst_list) {
/* If a pool is already handling a rebuild operation,
* wait to start the next operation until the current
* one completes
*/
if (pool_is_rebuilding(task->dst_pool_uuid) ||
task->dst_schedule_time > now ||
task->dst_schedule_time == (uint64_t)-1) {
struct rebuild_task *head_task = task;

Expand Down Expand Up @@ -2261,7 +2267,7 @@ rebuild_ults(void *arg)
}

}
dss_sleep(0);
dss_sleep(100);
}

/* If there are still rebuild task in queue and running list, then
Expand Down
Loading