Now that all there is one tree per thread group, all thread groups will
start on the same server. To prevent that, just insert the servers in a
different order for each thread group.
When using the round-robin load balancer, the major source of contention
is the lbprm lock, that has to be held every time we pick a server.
To mitigate that, make it so there are one tree per thread-group, and
one lock per thread-group. That means we now have a lb_fwrr_per_tgrp
structure that will contain the two lb_fwrr_groups (active and backup) as well
as the lock to protect them in the per-thread lbprm struct, and all
fields in the struct server are now moved to the per-thread structure
too.
Those changes are mostly mechanical, and brings good performances
improvment, on a 64-cores AMD CPU, with 64 servers configured, we could
process about 620000 requests par second, and we now can process around
1400000 requests per second.
Move the "next_weight" outside of fwrr_group, and inside struct lb_fwrr
directly, one for the active servers, one for the backup servers.
We will soon have one fwrr_group per thread group, but next_weight will
be global to all of them.
For both proxies and servers, properly calculates queueslength, which is
the total number of element in each queues (as they currently are only
using one queue, it is equivalent to the number of element of that
queue), and use it instead of the queue's length.
The leastconn and roundrobin functions mention that the server's lock
must be held while this is not true at all and it is not used either.
The "first" algo doesn't mention anything about the need for locking,
so let's mention that it uses the lbprm lock.
It was previously a spinlock, and it happens that a number of LB algos
only lock it for lookups, without performing any modification. Let's
first turn it to an rwlock and w-lock it everywhere. This is strictly
identical.
It was carefully checked that every HA_SPIN_LOCK() was turned to
HA_RWLOCK_WRLOCK() and that HA_SPIN_UNLOCK() was turned to
HA_RWLOCK_WRUNLOCK() on this lock. _INIT and _DESTROY were updated too.
This patch fixes all the leftovers from the include cleanup campaign. There
were not that many (~400 entries in ~150 files) but it was definitely worth
doing it as it revealed a few duplicates.
extern struct dict server_name_dict was moved from the type file to the
main file. A handful of inlined functions were moved at the bottom of
the file. Call places were updated to use server-t.h when relevant, or
to simply drop the entry when not needed.
The files remained mostly unchanged since they were OK. However, half of
the users didn't need to include them, and about as many actually needed
to have it and used to find functions like srv_currently_usable() through
a long chain that broke when moving the file.
global.h was one of the messiest files, it has accumulated tons of
implicit dependencies and declares many globals that make almost all
other file include it. It managed to silence a dependency loop between
server.h and proxy.h by being well placed to pre-define the required
structs, forcing struct proxy and struct server to be forward-declared
in a significant number of files.
It was split in to, one which is the global struct definition and the
few macros and flags, and the rest containing the functions prototypes.
The UNIX_MAX_PATH definition was moved to compat.h.
This one used to be stored into debug.h but the debug tools got larger
and require a lot of other includes, which can't use BUG_ON() anymore
because of this. It does not make sense and instead this macro should
be placed into the lower includes and given its omnipresence, the best
solution is to create a new bug.h with the few surrounding macros needed
to trigger bugs and place assertions anywhere.
Another benefit is that it won't be required to add include <debug.h>
anymore to use BUG_ON, it will automatically be covered by api.h. No
less than 32 occurrences were dropped.
The FSM_PRINTF macro was dropped since not used at all anymore (probably
since 1.6 or so).
All files that were including one of the following include files have
been updated to only include haproxy/api.h or haproxy/api-t.h once instead:
- common/config.h
- common/compat.h
- common/compiler.h
- common/defaults.h
- common/initcall.h
- common/tools.h
The choice is simple: if the file only requires type definitions, it includes
api-t.h, otherwise it includes the full api.h.
In addition, in these files, explicit includes for inttypes.h and limits.h
were dropped since these are now covered by api.h and api-t.h.
No other change was performed, given that this patch is large and
affects 201 files. At least one (tools.h) was already freestanding and
didn't get the new one added.
This is where other imported components are located. All files which
used to directly include ebtree were touched to update their include
path so that "import/" is now prefixed before the ebtree-related files.
The ebtree.h file was slightly adjusted to read compiler.h from the
common/ subdirectory (this is the only change).
A build issue was encountered when eb32sctree.h is loaded before
eb32tree.h because only the former checks for the latter before
defining type u32. This was addressed by adding the reverse ifdef
in eb32tree.h.
No further cleanup was done yet in order to keep changes minimal.
In fwrr_get_next_server(), we optionally pass a server to avoid. It
usually points to the current server during a redispatch operation. If
this server is usable, an "avoided" pointer is set and we continue to
look for another server. If in the end no other server is found, then
we fall back to this avoided one, which is still better than nothing.
The problem that may arise with threads is that in the mean time, this
avoided server might have received extra connections and might not be
usable anymore. This causes it to be queued a second time in the "full"
list and the loop to search for a server again, ending up on this one
again and so on.
This patch makes sure that we break out of the loop when we have to
pick the avoided server. It's probably what the code intended to do
as the current break statement causes fwrr_update_position() and
fwrr_dequeue_srv() to be called again on the avoided server.
It must be backported to 1.9 and 1.8, and seems appropriate for older
versions though it's unclear what the impact of this bug might be
there since the race doesn't exist and we're left with the double
update of the server's position.
An occasional divide by zero in the round-robin scheduler was addressed
in commit 9df86f997 ("BUG/MAJOR: lb/threads: fix insufficient locking on
round-robin LB") by grabing the server's lock in fwrr_get_server_from_group().
But it happens that this is not the correct approach as it introduces a
case of AB/BA deadlock reported by Maksim Kupriianov. This happens when
a server weight changes from/to zero while another thread extracts this
server from the tree. The reason is that the functions used to manipulate
the state work under the server's lock and grab the LB lock while the ones
used in LB work under the LB lock and grab the server's lock when needed.
This commit mostly reverts the changes above and instead further completes
the locking analysis performed on this code to identify areas that really
need to be protected by the server's lock, since this is the only algorithm
which happens to have this requirement. This audit showed that in fact all
locations which require the server's lock are already protected by the LB
lock. This was not noticed the first time due to the server's lock being
taken instead and due to some functions misleadingly using atomic ops to
modify server fields which are under the LB lock protection (these ones
were now removed).
The change consists in not taking the server's lock anymore here, and
instead making sure that the aforementioned function which used to
suffer from the server's weight becoming zero only uses a copy of the
weight which was preliminary verified to be non-null (when the weight
is null, the server will be removed from the tree anyway so there is
no need to recalculate its position).
With this change, the code survived an injection at 200k req/s split
on two servers with weights changing 50 times a second.
This commit must be backported to 1.9 only.
Maksim Kupriianov reported very strange crashes in fwrr_update_position()
which didn't make sense because of an apparent divide overflow except that
the value was not null in the core.
It happens that while the locking is correct in all the functions' call
graph, the uppermost one (fwrr_get_next_server()) incorrectly expected
that its target server was already locked when called. This stupid
assumption causd the server lock not to be held when calling the other
ones, explaining how it was possible to change the server's eweight by
calling srv_lb_commit_status() under the server lock yet collide with
its unprotected usage.
This commit makes sure that fwrr_get_server_from_group() retrieves a
locked server and that fwrr_get_next_server() is responsible for
unlocking the server before returning it. There is one subtlety in
this function which is that it builds a list of avoided servers that
were full while scanning the tree, and all of them are queued in a
full state so they must be unlocked upon return.
Many thanks to Maksim for providing detailed info allowing to narrow
down this bug.
This fix must be backported to 1.9. In 1.8 the lock seems much wider
and changes to the server's state are performed under the rendez-vous
point so this it doesn't seem possible that it happens there.
Since commit 3ff577e ("MAJOR: server: make server state changes
synchronous again"), srv_update_status() calls the various maintenance
operations of the LB algorithms (->set_server_up, ->set_server_down,
->update_server_weight()). These ones are called with a single thread
guaranteed by the rendez-vous point, so the fact that they're lacking
some locks has no effect. However we'll need to remove the rendez-vous
point so we have to take care of properly locking all the LB algos.
The comments have been properly updated on the various functions to
mention their locking expectations. All these functions are called
with the server lock held, and all of them now support concurrent
calls by using the lbprm's lock.
This fix doesn't need to be backported at the moment, though if any
check-specific issue surfaced in 1.8, it could make sense to reuse it.
A lock for LB parameters has been added inside the proxy structure and atomic
operations have been used to update server variables releated to lb.
The only significant change is about lb_map. Because the servers status are
updated in the sync-point, we can call recalc_server_map function synchronously
in map_set_server_status_up/down function.
The server state and weight was reworked to handle
"pending" values updated by checks/CLI/LUA/agent.
These values are commited to be propagated to the
LB stack.
In further dev related to multi-thread, the commit
will be handled into a sync point.
Pending values are named using the prefix 'next_'
Current values used by the LB stack are named 'cur_'
Till now, the server's state and flags were all saved as a single bit
field. It causes some difficulties because we'd like to have an enum
for the state and separate flags.
This commit starts by splitting them in two distinct fields. The first
one is srv->state (with its counter-part srv->prev_state) which are now
enums, but which still contain bits (SRV_STF_*).
The flags now lie in their own field (srv->flags).
The function srv_is_usable() was updated to use the enum as input, since
it already used to deal only with the state.
Note that currently, the maintenance mode is still in the state for
simplicity, but it must move as well.
We used to call srv_is_usable() with either the current state and weights
or the previous ones. This causes trouble for future changes, so let's first
split it in two variants :
- srv_is_usable(srv) considers the current status
- srv_was_usable(srv) considers the previous status
Detecting that a server's status has changed is a bit messy, as well
as it is to commit the status changes. We'll have to add new conditions
soon and we'd better avoid to multiply the number of touched locations
with the high risk of forgetting them.
This commit introduces :
- srv_lb_status_changed() to report if the status changed from the
previously committed one ;
- svr_lb_commit_status() to commit the current status
The function is now used by all load-balancing algorithms.
A crash was reported by Igor at owind when changing a server's weight
on the CLI. Lukas Tribus could reproduce a related bug where setting
a server's weight would result in the new weight being multiplied by
the initial one. The two bugs are the same.
The incorrect weight calculation results in the total farm weight being
larger than what was initially allocated, causing the map index to be out
of bounds on some hashes. It's easy to reproduce using "balance url_param"
with a variable param, or with "balance static-rr".
It appears that the calculation is made at many places and is not always
right and not always wrong the same way. Thus, this patch introduces a
new function "server_recalc_eweight()" which is dedicated to this task
of computing ->eweight from many other elements including uweight and
current time (for slowstart), and all users now switch to use this
function.
The patch is a bit large but the code was not trivially fixable in a way
that could guarantee this situation would not occur anymore. The fix is
much more readable and has been verified to work with all algorithms,
with both consistent and map-based hashes, and even with static-rr.
Slowstart was tested as well, just like enable/disable server.
The same bug is very likely present in 1.4 as well, so the patch will
probably need to be backported eventhough it will not apply as-is.
Thanks to Lukas and Igor for the information they provided to reproduce it.
The max weight of server is 256 now, but SRV_UWGHT_MAX is still 255. As a result,
FWRR will not work well when server's weight is 256. The description is as below:
There are some macros related to server's weight in include/types/server.h:
#define SRV_UWGHT_RANGE 256
#define SRV_UWGHT_MAX (SRV_UWGHT_RANGE - 1)
#define SRV_EWGHT_MAX (SRV_UWGHT_MAX * BE_WEIGHT_SCALE)
Since weight of server can be reach to 256 and BE_WEIGHT_SCALE equals to 16,
the max eweight of server should be 256*16 = 4096, it will exceed SRV_EWGHT_MAX
which equals to SRV_UWGHT_MAX*BE_WEIGHT_SCALE = 255*16 = 4080. When a server
with weight 256 is insterted into FWRR tree during initialization, the key value
of this server should be SRV_EWGHT_MAX - s->eweight = 4080 - 4096 = -16 which
is closed to UINT_MAX in unsigned type, so the server with highest weight will
be not elected as the first server to process request.
In addition, it is a better choice to compare with SRV_UWGHT_MAX than a magic
number 256 while doing check for the weight. The max number of servers for
round-robin algorithm is also updated.
Signed-off-by: Godbach <nylzhaowei@gmail.com>
All files referencing the previous ebtree code were changed to point
to the new one in the ebtree directory. A makefile variable (EBTREE_DIR)
is also available to use files from another directory.
The ability to build the libebtree library temporarily remains disabled
because it can have an impact on some existing toolchains and does not
appear worth it in the medium term if we add support for multi-criteria
stickiness for instance.
It was becoming painful to have all the LB algos in backend.c.
Let's move them to their own files. A few hashing functions still
need be broken in two parts, one for the contents and one for the
map position.
