ellis wrote:

iczfirz wrote:

 
 ip nat translation timeout 600 
 ip nat translation tcp-timeout 600 
 ip nat translation udp-timeout 15 
 ip nat translation dns-timeout 600 
 ip nat translation port-timeout tcp 22 86400 
 ip nat translation port-timeout tcp 23 3600 
 ip nat translation port-timeout tcp 3223 3600 
 ip nat translation max-entries 1000 
 ip nat pool RR-1 1.0.0.31 1.0.0.31 netmask 255.255.255.0 type rotary 
 ip nat pool IP-30 202.123.165.30 202.123.165.30 netmask 255.255.255.248 
 ip nat inside source list 5 pool IP-30 overload 
 ip nat inside destination list 111 pool RR-1 
 ! 
 ip classless 
 ip route 0.0.0.0 0.0.0.0 202.123.165.25 
 ! 
 access-list 5 permit 1.0.0.0 0.0.0.255 
 access-list 111 permit tcp any host 202.123.165.28 eq www 
 access-list 111 permit tcp any host 202.123.165.28 eq  
 

RR-1暫時只有一部server.
server bootup IP: 1.0.0.10
sevice (alias) IP: 1.0.0.31
古怪左部router行唔多個一日就會自己除時唔曉再做nat load balancing. 但nat overload就仲做到.. 點解? 當做唔到load balancing時我比條 ip nat inside source static 1.0.0.31 202.123.165.28 extendable 佢再即時no番佢佢又會再行番.. 但都係過唔到一日...

我諗不如唔好行 ip nat translation o個 d timeout option 先, 因為會影響
dynamic translation.

iczfirz wrote:
但偉兄你有冇見過有人會行overload+rotary? 我係internet找左好耐都冇人講...

重有... 似乎係我係server用左ip alias先會出現問題.. 如果唔用alias既 ip 就冇乜事

我開頭試過唔用ip nat tran.. 但因為呢度有好多http同mysql connection.. 隻router會因為有太多nat translation行到死.. 唉... 不知點算好..