I have pureftp running on an AWS ec2 instance. I'm trying to get it to run in passive mode which I thought was working, however I'm finding it may not be working correctly. I'm receiving the following error in FileZilla
Status: Connected
Status: Retrieving directory listing...
Status: Server sent passive reply with unroutable address. Using server address instead.
Status: Directory listing of "/" successful
The odd part is some people are unable to log in while others are.
I have the following pureftp configuration
Port Range
#Port range for passive connections replies. - for firewalling.
PassivePortRange `50000 50100`
PASV IP
#Force an IP address in PASV/EPSV/SPSV replies. - for NAT.
#Symbolic host names are also accepted for gateways with dynamic IP
#addresses.
ForcePassiveIP `ftp.mydomain.com` "my cname record is mapped to my ec2 public dns"
When I view the local port range on the server, /proc/sys/net/ipv4/ip_local_port_range the following are open.
32768 61000
My ec2 security group has port 50000 - 50100 open
When I view my server logs I don't see much other than this every once in a while.
Feb 5 08:57:41 ip-172-11-42-52 dhclient[1062]: DHCPREQUEST on eth0 to 172.11.32.1 port 67 (xid=0x601547fd)
Feb 5 08:57:41 ip-172-11-42-52 dhclient[1062]: DHCPACK from 172.11.32.1 (xid=0x601547fd)
Feb 5 08:57:43 ip-172-11-42-52 dhclient[1062]: bound to 172.11.42.52 -- renewal in 1417 seconds.
Anybody have any idea where things might be going wrong?
So I believe I resolved this issue. It appears as if aws was dynamically changing ip addresses. When I was referencing ftp.domain.com for my passiveip the ip that it resolved to didn't match the initial ip tied to the cname record.
The solution was to assign a static elastic ip to my ec2 instance and set my passiveip in pureftp to my static elastic ip. Thus far it appears to have resolved my issue.