|
Hello ratz,
Monday, November 20, 2000, 5:17:30 PM, you wrote:
r> Sorry for the late reply but I took off this weekend.
Never mind.
r> That's basically the approach. Since you're suggesting to load balance
r> http-traffic, this is extremely easy. (Hope I correctly understand your
r> problem) Take 2 or more realserver with the same DocumentRoot and the
r> same ../cgi-bin/. The synchronisation can be done via rsync and if you
r> want to have it secure rsync -e ssh ... . Since we can assume you don't
r> change your static webserver content too often and you also don't change
r> your cgi's you can do a cronjob rsyncing the webservers let's say every
r> three hours. This poor-man-but-extremely-well-working solution has proven
r> very reliable in productionary envirement for mid-range webserver-farms.
So if i have dynamic webserver content, the realservers will share one
database or they each have one and keep them snychro? If those
requests are mostly database related, i guess the workload will not be
balanced very well if there is only one database. And if those
realsevers each have a database, let's say MySQL, is it possible to
keep them synchro?
r> Ok, the second thing you'd like to know is how to check the service
r> running on a realserver (please correct me if I'm wrong): I call this
r> healthchecking (commercial loadbalancer vendors tend to call it alike)
r> What you do here is: The loadbalancer runs a daemon that checks periodi-
r> cally all realservers services like for example http, telnet. This check
r> can be from poor-man solution (icmp) to highend (connect to webserver,
r> start a request to the database and verify it's content). Your daemon
r> then performes those checks agains all realservers and if a check fails
r> the realserver will be taken out. The check goes on and if the server
r> or service comes back, it will be inserted again into the table with
r> ipvsadm. NOTE: This works EXTREMELY well!!! Only negative aspect: You
r> have to write your own daemon. LVS doesn't provide this for several
r> reasons.
Sorry, i think i did not make myself very clear in last letter, but
your answer is very helpful, i can learn a lot from it as well :-)
What i really wanted to say is how to verify those scheduling
algorithms like rr,wrr,lc,wlc, using browser? But cache is a big
problem when making verifications.
r> Hope this answers your questions:
r> Best regards,
r> Roberto Nibali, ratz
It's so kind of you!
Your answers are very helpful.
Thank you very much!
--
Best regards,
huatao mailto:thua@xxxxxxxxxxxxxxx
|
