mirror of
http://git.haproxy.org/git/haproxy.git/
synced 2025-03-20 09:56:47 +00:00
f76e6cad83
When 'minconn' is set, the number of simultaneous sessions sent to the server will be limited by a dynamic value depending on the global load on the instance itself. The principle is to fix the maximal concurrency on the server proportionnally to the instance's usage relative to its maxconn, with a minimum fixed to <minconn>. The formula for the number of simultaneous sessions sent to the server is then max(srv_minconn, srv_maxconn*px_conn/px_maxconn). This helps unloading the servers when the load is very low.
56 lines
1.7 KiB
Plaintext
56 lines
1.7 KiB
Plaintext
'+' = done, '-' = todo, '*' = done except doc
|
|
|
|
1.2.12 :
|
|
+ weighted RR/SH
|
|
|
|
1.2.13 :
|
|
+ maxconn
|
|
+ queueing
|
|
|
|
1.2.14 :
|
|
* HTML status page
|
|
|
|
stats enable
|
|
stats uri /?stats
|
|
stats realm w.ods.org\ statistics
|
|
stats auth user1:pass1
|
|
stats auth user2:pass2
|
|
stats auth user3:pass3
|
|
stats scope <px_id> | '.'
|
|
|
|
+ allow server-less proxies (for stats)
|
|
|
|
- separate timeout controls
|
|
|
|
- option 'abortonclose' : if the session is queued or being connecting
|
|
to the server, and the client sends a shutdown(), then decide to abort
|
|
the session early because in most situations, this will be caused by
|
|
a client hitting the 'Stop' button, so there's no reason to overload
|
|
the servers with unservable requests. However, this is not HTTP compliant
|
|
and might cause little trouble to some very specific clients used to
|
|
close immediately after sending the request (no support for KA, which ones?)
|
|
|
|
+ minconn : makes the server's maxconn dynamic, which will be computed as a
|
|
ratio of the proxy's sessions :
|
|
srv->effective_maxconn =
|
|
max(srv->maxconn * px->nbsess / px->maxconn, srv->minconn)
|
|
|
|
1.3 :
|
|
- handle half-closed connections better (cli/srv would not distinguish
|
|
DATA/SHUTR/SHUTW, it would be a session flag which would tell shutr/shutw).
|
|
Check how it got changed in httpterm.
|
|
|
|
- 3 memory models : failsafe (prealloc), normal (current), optimal (alloc on
|
|
demand)
|
|
|
|
- wait queues replaced for priority-based trees
|
|
|
|
- ability to assign a prio based on L7 matching
|
|
|
|
- prio-based O(1) scheduler
|
|
|
|
- maxconn reserve for VIP/admins
|
|
|
|
- verify if it would be worth implementing an epoll_ctl_batch() for Linux
|
|
|