Closed fruch closed 2 years ago
Please, leave only "keyspace" part of the change. For full workability we will need the unreleased version of
gocql
with the following fix: scylladb/gocql#104
I'm not sure how scylladb/gocql#104 is related to should we advance the version or not, why wouldn't we ?
if for testing specific fixes, we can build and test, i'm not following what do you mean here.
Please, leave only "keyspace" part of the change. For full workability we will need the unreleased version of
gocql
with the following fix: scylladb/gocql#104I'm not sure how scylladb/gocql#104 is related to should we advance the version or not, why wouldn't we ?
if for testing specific fixes, we can build and test, i'm not following what do you mean here.
At first, we will need the update after merge of the mentioned issue - it is for sure. At second, such a change must be in a separate PR, because both existing commits are not dependant on each other.
Please, leave only "keyspace" part of the change. For full workability we will need the unreleased version of
gocql
with the following fix: scylladb/gocql#104I'm not sure how scylladb/gocql#104 is related to should we advance the version or not, why wouldn't we ? if for testing specific fixes, we can build and test, i'm not following what do you mean here.
At first, we will need the update after merge of the mentioned issue - it is for sure. At second, such a change must be in a separate PR, because both existing commits are not dependant on each other.
do you have any evidence this specific issue, part of what holding the retries
Please, leave only "keyspace" part of the change. For full workability we will need the unreleased version of
gocql
with the following fix: scylladb/gocql#104I'm not sure how scylladb/gocql#104 is related to should we advance the version or not, why wouldn't we ? if for testing specific fixes, we can build and test, i'm not following what do you mean here.
At first, we will need the update after merge of the mentioned issue - it is for sure. At second, such a change must be in a separate PR, because both existing commits are not dependant on each other.
do you have any evidence this specific issue, part of what holding the retries
Seem what's described in https://github.com/gocql/gocql/issues/1621 might be affecting us, and making scylla-bench send unblanced traffic to scylla
REF: https://github.com/gocql/gocql/issues/1621