Can not failover up to good server if one server down with consistent_hash method

[xiaoxuanzi]

Hi: I found a error: upstream test have two servers:

upstream test{ server 192.168.46.111:80; server 192.168.46.110:80; }

if upstream use consistent_hash method, if one of upstream server is down, this method can not failover up to good server.

I guess Problem may be in the function of try_cluster_consistent_hash:

_
-- try backup node local hash_backup_node = cls.hash_backup_node or 1 local q = (p + hash % hash_backup_node + 1) % server_len + 1 if p ~= q then local try = cls.try or #cls.servers res, err = base.tryserver(skey, ups, cls.servers[q], callback, args, try - 1) if res then return res end end

It so?

I am anticipating your reply.

[yejingx]

Please upgrade lua-resty-checkups to the newest version and try again. This problem seems to have been fixed in the new version.

Related issue: https://github.com/upyun/lua-resty-checkups/issues/3

[xiaoxuanzi]

thanks a lot