Due to weight being set to 0 for down peers, order of peers after sorting
wasn't the same as without the "down" flag (with down peers at the end),
resulting in client rebalancing for clients on other servers. The only
rebalancing which should happen after adding "down" to a server is one
for clients on the server.
The problem was introduced in r1377 (which fixed endless loop by setting
weight to 0 for down servers). The loop is no longer possible with new
smooth algorithm, so preserving original weight is safe.
For edge case weights like { 5, 1, 1 } we now produce { a, a, b, a, c, a, a }
sequence instead of { c, b, a, a, a, a, a } produced previously.
Algorithm is as follows: on each peer selection we increase current_weight
of each eligible peer by its weight, select peer with greatest current_weight
and reduce its current_weight by total number of weight points distributed
among peers.
In case of { 5, 1, 1 } weights this gives the following sequence of
current_weight's:
a b c
0 0 0 (initial state)
5 1 1 (a selected)
-2 1 1
3 2 2 (a selected)
-4 2 2
1 3 3 (b selected)
1 -4 3
6 -3 4 (a selected)
-1 -3 4
4 -2 5 (c selected)
4 -2 -2
9 -1 -1 (a selected)
2 -1 -1
7 0 0 (a selected)
0 0 0
To preserve weight reduction in case of failures the effective_weight
variable was introduced, which usually matches peer's weight, but is
reduced temporarily on peer failures.
This change also fixes loop with backup servers and proxy_next_upstream
http_404 (ticket #47), and skipping alive upstreams in some cases if there
are multiple dead ones (ticket #64).
Such upstreams cause CPU hog later in the code as number of peers isn't
expected to be 0. Currently this may happen either if there are only backup
servers defined in an upstream block, or if server with ipv6 address used
in an upstream block.
Previously nginx used to mark backend again as live as soon as fail_timeout
passes (10s by default) since last failure. On the other hand, detecting
dead backend takes up to 60s (proxy_connect_timeout) in typical situation
"backend is down and doesn't respond to any packets". This resulted in
suboptimal behaviour in the above situation (up to 23% of requests were
directed to dead backend with default settings).
More detailed description of the problem may be found here (in Russian):
http://mailman.nginx.org/pipermail/nginx-ru/2011-August/042172.html
Fix is to only allow one request after fail_timeout passes, and
mark backend as "live" only if this request succeeds.
Note that with new code backend will not be marked "live" unless "check"
request is completed, and this may take a while in some specific workloads
(e.g. streaming). This is believed to be acceptable.
The following configuration causes nginx to hog cpu due to infinite loop
in ngx_http_upstream_get_peer():
upstream backend {
server 127.0.0.1:8080 down;
server 127.0.0.1:8080 down;
}
server {
...
location / {
proxy_pass http://backend;
}
}
Make sure we don't loop infinitely in ngx_http_upstream_get_peer() but stop
after resetting peer weights once.
Return 0 if we are stuck. This is guaranteed to work as peer 0 always exists,
and eventually ngx_http_upstream_get_round_robin_peer() will do the right
thing falling back to backup servers or returning NGX_BUSY.
by ngx_http_upstream_create_round_robin_peer(), since the peer lives
only during request so the saved SSL session will never be used again
and just causes memory leak
patch by Maxim Dounin
