replication-manager
replication-manager copied to clipboard
extproxy support missing ?
Hi,
I'm trying to use extproxy and extproxy-address but when I read the code I have the impression that it's actually missing.
More precisely, it fails here:
replication-manager-osc[16620]: [signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0xa8f21b]
replication-manager-osc[16620]: goroutine 16 [running]:
replication-manager-osc[16620]: github.com/signal18/replication-manager/cluster.(*Cluster).initProxies(0xc4203aa000)
replication-manager-osc[16620]: #011/var/jenkins_home/workspace/go/src/github.com/signal18/replication-manager/cluster/prx.go:385 +0x9b```
Hi Thanks
What replication-manager version are you using ? ./replication-manager-osc monitor --help | grep extp --extproxy External proxy can be used to specify a route manage with external scripts --extproxy-address string Network address when route is manage via external script, host:[port] format
parameter in config should be extproxy = true extproxy-address ="ip:port" to join the cluster
I'm using the 2.0 package for debian from your repo :
replication-manager-osc_2.0.1_amd64.deb
# replication-manager-osc monitor --help | grep extp
--extproxy External proxy can be used to specify a route manage with external scripts
--extproxy-address string Network address when route is manage via external script, host:[port] format
When I read the initProxies (and others) I don't see how it could work actually since the ExtProxyOn and ExtProxyVIP are not used there, but maybe I'm wrong.
Also can you contact me to talk professional services and licensing of the Pro version ? [email protected]
I can see it from https://github.com/signal18/replication-manager/blob/2.0/cluster/prx.go line 157 to 165
This extproxy is used by replication-manager to externally join the master, mainly in pseudo gtid injection or benchmarking also for extra tools like sphinx indexer to know how to query the database to fetches indexes If you get a load balancer or cloud Elastic IP or failover proxies !
Indeed, it's very strange because if I take the line of the crash it's
cluster.LogPrintf(LvlInfo, "Init Proxy Type: %s Host: %s Port: %s", pr.Type, pr.Host, pr.Port)
which doesn't make sense (the values I set are :
extproxy = true
extproxy-address = "127.0.0.1:3306"
In the file you send me there is extproxy-address = “127.0.0.1:3306" instead of extproxy-address = "127.0.0.1:3306"
See first quote is broken hoping it's a copy past issue and not in original conf file , can you confirm , i can not reproduce the issue after fixing this !
Hello,
Can you please try install last release from our ci https://ci.signal18.io:9443/ or depot https://signal18.io/downloads
Also i'm intrested in getting the full log file or by running replication-manager by hand and sending stdout or by setting log-file="balbla" in config file