Skip to content
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

DAOS-16005 object: check resent coll_punch on leader and relay engine #14659

Merged
merged 1 commit into from
Jul 22, 2024
Merged

DAOS-16005 object: check resent coll_punch on leader and relay engine #14659

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
5 changes: 4 additions & 1 deletion src/dtx/dtx_common.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1274,7 +1274,7 @@ dtx_leader_end(struct dtx_leader_handle *dlh, struct ds_cont_hdl *coh, int resul

dtx_shares_fini(dth);

if (daos_is_zero_dti(&dth->dth_xid) || unlikely(result == -DER_ALREADY) || dlh->dlh_relay)
if (daos_is_zero_dti(&dth->dth_xid) || unlikely(result == -DER_ALREADY))
goto out;

if (unlikely(coh->sch_closed)) {
Expand Down Expand Up @@ -1328,6 +1328,9 @@ dtx_leader_end(struct dtx_leader_handle *dlh, struct ds_cont_hdl *coh, int resul
D_ASSERTF(0, "Unexpected DTX "DF_DTI" status %d\n", DP_DTI(&dth->dth_xid), status);
}

if (dlh->dlh_relay)
goto out;

/*
* Even if the transaction modifies nothing locally, we still need to store
* it persistently. Otherwise, the subsequent DTX resync may not find it as
Expand Down
6 changes: 5 additions & 1 deletion src/object/srv_obj.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3568,8 +3568,12 @@ obj_tgt_punch(struct obj_tgt_punch_args *otpa, uint32_t *shards, uint32_t count)
goto out;
}

if (dth != NULL)
if (dth != NULL) {
if (dth->dth_prepared)
D_GOTO(out, rc = 0);

goto exec;
}

if (opi->opi_flags & ORF_RESEND) {
tmp = opi->opi_epoch;
Expand Down
28 changes: 19 additions & 9 deletions src/vos/vos_dtx.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -645,7 +645,8 @@ dtx_rec_release(struct vos_container *cont, struct vos_dtx_act_ent *dae,
dbd = dae->dae_dbd;
dae_df = umem_off2ptr(umm, dae->dae_df_off);

D_ASSERT(dae_df != NULL);
D_ASSERTF(dae_df != NULL, "Hit invalid DTX entry "DF_DTI" when release for %s\n",
DP_DTI(&DAE_XID(dae)), abort ? "abort" : "commit");
D_ASSERTF(dbd->dbd_magic == DTX_ACT_BLOB_MAGIC,
"Invalid blob %p magic %x for "DF_DTI" (lid %x)\n",
dbd, dbd->dbd_magic, DP_DTI(&DAE_XID(dae)), DAE_LID(dae));
Expand Down Expand Up @@ -1177,13 +1178,16 @@ vos_dtx_check_availability(daos_handle_t coh, uint32_t entry,
}

if (intent == DAOS_INTENT_PURGE) {
uint32_t age = d_hlc_age2sec(DAE_XID(dae).dti_hlc);

/*
* The DTX entry still references related data record,
* then we cannot (vos) aggregate related data record.
*/
if (d_hlc_age2sec(DAE_XID(dae).dti_hlc) >= DAOS_AGG_THRESHOLD)
D_WARN("DTX "DF_DTI" (%u) still references the data, cannot be (VOS) "
"aggregated\n", DP_DTI(&DAE_XID(dae)), vos_dtx_status(dae));
if (age >= DAOS_AGG_THRESHOLD)
D_WARN("DTX "DF_DTI" (state:%u, age:%u) still references the data, "
"cannot be (VOS) aggregated\n",
DP_DTI(&DAE_XID(dae)), vos_dtx_status(dae), age);
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's better to rate limit this warning message, otherwise, the log file will be flooded with such message when some data failed to be DTX committed for a long time.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The warning message may be triggered for different DTX entries. Limiting the rate may miss some DTX information? Not sure whether we have some efficient way to skip repeated log message or not.


return ALB_AVAILABLE_DIRTY;
}
Expand Down Expand Up @@ -1420,21 +1424,25 @@ vos_dtx_validation(struct dtx_handle *dth)
if (rc == -DER_NONEXIST) {
rc = dbtree_lookup(cont->vc_dtx_committed_hdl, &kiov, &riov);
if (rc == 0)
return DTX_ST_COMMITTED;
D_GOTO(out, rc = DTX_ST_COMMITTED);
}

/* Failed to lookup DTX entry, in spite of whether it is DER_NONEXIST
* or not, then handle it as aborted that will cause client to retry.
*/
return DTX_ST_ABORTED;
D_GOTO(out, rc = DTX_ST_ABORTED);
}

dae = riov.iov_buf;
} else if (unlikely(dae == NULL)) {
return DTX_ST_COMMITTED;
D_GOTO(out, rc = DTX_ST_COMMITTED);
}

return vos_dtx_status(dae);
rc = vos_dtx_status(dae);

out:
dth->dth_need_validation = 0;
return rc;
}

static int
Expand Down Expand Up @@ -2934,8 +2942,10 @@ vos_dtx_cleanup_internal(struct dtx_handle *dth)
/* Only keep the DTX entry (header) for handling resend RPC,
* remove DTX records, purge related VOS objects from cache.
*/
if (dae != NULL)
if (dae != NULL) {
D_ASSERT(!vos_dae_is_prepare(dae));
dtx_act_ent_cleanup(cont, dae, dth, true);
}
} else {
d_iov_set(&kiov, &dth->dth_xid, sizeof(dth->dth_xid));
d_iov_set(&riov, NULL, 0);
Expand Down
Loading