You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We have a situation where phpredis starts failing with errors
"RedisCluster::__construct(): php_network_getaddresses: getaddrinfo failed: Name or service not known"
and "Fatal error: Uncaught RedisClusterException: Couldn't map cluster keyspace using any provided seed"
When there is a large amount of requests on a cluster with 20 or more nodes.
We use AWS Elasticache as our redis cluster. When the node count on the cluster is below 20 everything works fine, even under high loads. Once the node count is 20 or more we start seeing the above errors.
We use a DNS address(provided by Elasticache) as the host endpoint and not individually specified hosts.
We know this is a phpredis issue because the same redis cluster is being used by go-based clients as well and there we see no issues(no errors or slowness)
We have tried enabling redis.clusters.cache_slots = 1 but that did not help
Tested latest phpredis(6.0.2) as well with no help
I suspect it has something to do with the logic behind the cluster nodes connections
Any help would be appreciated!
Thanks!
Expected behavior
No errors
Actual behavior
Getting the errors above
I'm seeing this behavior on
OS: Centos 7
Redis: 6.2.6
PHP: 7.4.3
phpredis: 5.1.1 / 6.0.2
Steps to reproduce, backtrace or example script
Redis cluster with 20 nodes or more.
High load simulation, eg 1000 concurrent connections creating 1000 random read write requests
I can provide the php script i was creating the artificial load with if needed.
I've checked
There is no similar issue from other users
Issue isn't fixed in develop branch
The text was updated successfully, but these errors were encountered:
It's probably going to be tough to track this down without being able to replicate it, but the error indicates something is failing at the kernel/OS layer.
php_network_getaddresses: getaddrinfo failed: Name or service not known
That means that the getaddrinfo syscall is not able to resolve the host. Perhaps there is some sort of resource exhaustion going on here that is causing these symptoms?
Hi
We have a situation where phpredis starts failing with errors
"RedisCluster::__construct(): php_network_getaddresses: getaddrinfo failed: Name or service not known"
and "Fatal error: Uncaught RedisClusterException: Couldn't map cluster keyspace using any provided seed"
When there is a large amount of requests on a cluster with 20 or more nodes.
We use AWS Elasticache as our redis cluster. When the node count on the cluster is below 20 everything works fine, even under high loads. Once the node count is 20 or more we start seeing the above errors.
We use a DNS address(provided by Elasticache) as the host endpoint and not individually specified hosts.
We know this is a phpredis issue because the same redis cluster is being used by go-based clients as well and there we see no issues(no errors or slowness)
We have tried enabling
redis.clusters.cache_slots = 1
but that did not helpTested latest phpredis(6.0.2) as well with no help
I suspect it has something to do with the logic behind the cluster nodes connections
Any help would be appreciated!
Thanks!
Expected behavior
No errors
Actual behavior
Getting the errors above
I'm seeing this behavior on
Steps to reproduce, backtrace or example script
Redis cluster with 20 nodes or more.
High load simulation, eg 1000 concurrent connections creating 1000 random read write requests
I can provide the php script i was creating the artificial load with if needed.
I've checked
develop
branchThe text was updated successfully, but these errors were encountered: