nsinnovations/erpnext
2
0
Fork 0
mirror of https://github.com/frappe/erpnext.git synced 2026-09-17 10:36:31 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(stock): fall back gracefully when transaction_advisory_lock is unavailable

Browse Source 
Same hasattr pattern as repost_gate: an ERPNext ahead of its frappe build keeps
the status-quo serialization-failure retries instead of failing every stock
submission on postgres.
This commit is contained in:
Mihir Kandoi
2026-07-16 09:32:03 +05:30
parent 35a9d7b09c
commit 897eca895a
1 changed files with 4 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
erpnext/stock/stock_ledger.py
View File

@@ -294,8 +294,10 @@ def sle_processing_gate(item_code, warehouse):
gap locks its previous-SLE locking reads take (which also block, then reveal, concurrent gap locks its previous-SLE locking reads take (which also block, then reveal, concurrent
inserts); postgres locking reads never see rows another transaction is inserting, so without inserts); postgres locking reads never see rows another transaction is inserting, so without
this gate two concurrent writers compute from the same stale previous SLE and the loser's Bin this gate two concurrent writers compute from the same stale previous SLE and the loser's Bin
write is lost. Txn-scoped and re-entrant; released at commit/rollback.""" write is lost. Txn-scoped and re-entrant; released at commit/rollback. hasattr keeps a frappe
if frappe.db.db_type == "postgres": predating transaction_advisory_lock on the status quo (serialization-failure retries) instead
of breaking every stock submission."""
if frappe.db.db_type == "postgres" and hasattr(frappe.db, "transaction_advisory_lock"):
frappe.db.transaction_advisory_lock(("stock-sle", item_code, warehouse), timeout=REPOST_LOCK_TIMEOUT) frappe.db.transaction_advisory_lock(("stock-sle", item_code, warehouse), timeout=REPOST_LOCK_TIMEOUT)