193 reputation
18
bio website
location Singapore
age 22
visits member for 3 years
seen Jun 15 at 21:50

Autodidact of various trades, geek of several.





There are probably two types of coders; the one who took up this craft before he knows it’s a quadzillion dollar industry, and the one after.






Sold as cheap, forced labour for 16056/16056 hours. When the people fear the government, there is tyranny. When governments fear the people, there is liberty.























Postal: PacerierSE Gmail


May
22
comment Referer is passed from HTTPS to HTTP in some cases… How?
Then it isn't violating the RFC is it?
May
14
comment Referer is passed from HTTPS to HTTP in some cases… How?
"Should" means optional right?
May
14
comment Why is DNS failover not recommended?
Training your users this way... Doesn't this make them more liable to getting phished?
May
14
comment Why is DNS failover not recommended?
@Ryan, What do you mean when you say "ghetto"?
May
14
comment Multiple data centers and HTTP traffic: DNS Round Robin is the ONLY way to assure instant fail-over?
@JesperMortensen, When you say 'intelligent' DNS, do you mean split-horizon DNS? Or do you mean something else (deciding based on factors beyond source IP)?
May
14
comment Multiple data centers and HTTP traffic: DNS Round Robin is the ONLY way to assure instant fail-over?
@vmiazzo, Your pdf link is down... Do you mean 15 minutes or 20 seconds to 15 minutes?
May
14
comment Why is DNS failover not recommended?
Cian's answer serverfault.com/a/60562/87017 directly contradicts your one..... so who is right?
May
14
comment Why is DNS failover not recommended?
Also see serverfault.com/q/315199/87017
May
14
comment Is round-robin DNS a possible solution for high availability?
@RyanRies, Does the RFC give any recommendations with regards to client behavior? Is Telnet compliant when it tries out all the IPs instead of merely using the first one?
May
14
comment How could dynamic DNS work if DNS updates take hours to propagate?
How bad is the situation in the wild? I'm referring to rogue DNS servers which doesn't respect TTLs. Since most companies don't write their own DNS software, do you know of any DNS software (djbdns, BIND, etc) that tries to play punk?
May
14
revised How could dynamic DNS work if DNS updates take hours to propagate?
deleted 16 characters in body; edited title
May
13
comment DNS servers already use anycast. Will adding more IPs enhance scalability?
@kasperd, What about special reserved IP addresses which belongs to no one?
May
13
revised DNS servers already use anycast. Will adding more IPs enhance scalability?
added 30 characters in body
May
13
revised DNS servers already use anycast. Will adding more IPs enhance scalability?
edited tags
May
13
comment DNS servers already use anycast. Will adding more IPs enhance scalability?
@kasperd, Would it be possible to dynamically change the 5 IP addresses in our DNS reply? For example, we can observe that the DNS request is coming from Asia via it's IP, then we customize a DNS reply to return the IP addresses of 5 of our Asia servers. If we see the DNS request coming from USA, we customize our DNS reply to return the IP addresses of 5 of our USA servers. Will this approach work?
May
13
revised DNS servers already use anycast. Will adding more IPs enhance scalability?
added 54 characters in body
May
13
comment DNS servers already use anycast. Will adding more IPs enhance scalability?
@kasperd, Btw why do you say that using multiple unicast addresses does not achieve load balancing? Isn't it the case that the load is balanced out among this pool of unicast addresses? For example, the Asia client will be given the Asia server's unicast address and the US client will be given the US server's unicast address.
May
13
accepted DNS servers already use anycast. Will adding more IPs enhance scalability?
May
13
revised DNS servers already use anycast. Will adding more IPs enhance scalability?
added 12 characters in body; edited title
May
13
comment DNS servers already use anycast. Will adding more IPs enhance scalability?
Do you mean that the best way to achieve scalability is to employ many secondary unicast addresses (a.k.a the old way of setting ns1.domain, ns2.d, ns3.d ... ns300.d) instead of using anycast?