firestore: don't retry reads that fail with Aborted
For transaction reads that fail with ABORTED, we need to rollback and start a new transaction. Our current configuration makes it so that GAPIC retries ABORTED reads multiple times without making any progress. Instead, we should retry at the transaction level. PiperOrigin-RevId: 289532382
This commit is contained in:
parent
1dbfd3fe43
commit
9118db63d1
|
|
@ -35,7 +35,6 @@
|
|||
"maxBackoff": "60s",
|
||||
"backoffMultiplier": 1.3,
|
||||
"retryableStatusCodes": [
|
||||
"ABORTED",
|
||||
"UNAVAILABLE",
|
||||
"INTERNAL",
|
||||
"DEADLINE_EXCEEDED"
|
||||
|
|
@ -56,7 +55,6 @@
|
|||
"maxBackoff": "60s",
|
||||
"backoffMultiplier": 1.3,
|
||||
"retryableStatusCodes": [
|
||||
"ABORTED",
|
||||
"UNAVAILABLE",
|
||||
"INTERNAL",
|
||||
"DEADLINE_EXCEEDED"
|
||||
|
|
@ -106,7 +104,6 @@
|
|||
"maxBackoff": "60s",
|
||||
"backoffMultiplier": 1.3,
|
||||
"retryableStatusCodes": [
|
||||
"ABORTED",
|
||||
"UNAVAILABLE",
|
||||
"INTERNAL",
|
||||
"DEADLINE_EXCEEDED"
|
||||
|
|
|
|||
Loading…
Reference in New Issue