This commit is a direct follow-up of the previous one. It defines a new
server keyword check-pool-conn-name. It is used as the default value for
the name parameter of idle connection hash generation.
Its behavior is similar to server keyword pool-conn-name, but reserved
for checks reuse. If check-pool-conn-name is set, it is used in priority
to match a connection for reuse. If unset, a fallback is performed on
check-sni.
Define a new server keyword check-reuse-pool, and its counterpart with a
"no" prefix. For the moment, only parsing is implemented. The real
behavior adjustment will be implemented in the next patch.
The progressive adoption of OpenSSL 3 and its abysmal handshake
performance has started to reveal situations where it simply isn't
possible anymore to succesfully run health checks on many servers,
because between the moment all the checks are started and the moment
the handshake finally completes, the timeout has expired!
This also has consequences on production traffic which gets
significantly delayed as well, all that for lots of checks. While it's
possible to increase the check delays, it doesn't solve everything as
checks still take a huge amount of time to converge in such conditions.
Here we take a different approach by permitting to enforce the maximum
concurrent checks per thread limitation and implementing an ordered
queue. Thanks to this, if a thread about to start a check has reached
its limit, it will add the check at the end of a queue and it will be
processed once another check is finished. This proves to be extremely
efficient, with all checks completing in a reasonable amount of time
and not being disturbed by the rest of the traffic from other checks.
They're just cycling slower, but at the speed the machine can handle.
One must understand however that if some complex checks perform multiple
exchanges, they will take a check slot for all the required duration.
This is why the limit is not enforced by default.
Tests on SSL show that a limit of 5-50 checks per thread on local
servers gives excellent results already, so that could be a good starting
point.
The goal here is to explicitly mark that a check was migrated so that
we don't do it again. This will allow us to perform other actions on
the target thread while still knowing that we don't want to be migrated
again. The new READY bit combine with SLEEPING to form 4 possible states:
SLP RDY State Description
0 0 - (reserved)
0 1 RUNNING Check is bound to current thread and running
1 0 SLEEPING Check is sleeping, not bound to a thread
1 1 MIGRATING Check is migrating to another thread
Thus we set READY upon migration, and check for it before migrating, this
is sufficient to prevent a second migration. To make things a bit clearer,
the SLEEPING bit was switched with FASTINTER so that SLEEPING and READY are
adjacent.
The CHK_ST_SLEEPING state was introduced by commit d114f4a68 ("MEDIUM:
checks: spread the checks load over random threads") to indicate that
a check was not currently bound to a thread and that it could easily
be migrated to any other thread. However it did not start the checks
in this state, meaning that they were not redispatchable on startup.
Sometimes under heavy load (e.g. when using SSL checks with OpenSSL 3.0)
the cost of setting up new connections is so high that some threads may
experience connection timeouts on startup. In this case it's better if
they can transfer their excess load to other idle threads. By just
marking the check as sleeping upon startup, we can do this and
significantly reduce the number of failed initial checks.
With previous commit, 9e080bf ("BUG/MINOR: checks: make sure fastinter is used
even on forced transitions"), on-error mark-down|sudden-death|fail-check are
now working as expected.
However, on-error fastinter remains broken because srv_getinter(), used in
the above commit to check the expiration date, won't return fastinter interval
if server health is maxed out (which is the case with on-error fastinter mode).
To fix this, we introduce a check flag named CHK_ST_FASTINTER.
This flag is set when on-error is triggered. This way we can force
srv_getinter() to return fastinter interval whenever the flag is set.
The flag is automatically cleared as soon as the new check task expiry is
recalculated in process_chk_conn().
This restores original behavior prior to d114f4a ("MEDIUM: checks: spread the
checks load over random threads").
It must be backported to 2.7 along with the aforementioned commits.
The CPU usage pattern was found to be high (5%) on a machine with
48 threads and only 100 servers checked every second That was
supposed to be only 100 connections per second, which should be very
cheap. It was figured that due to the check tasks unbinding from any
thread when going back to sleep, they're queued into the shared queue.
Not only this requires to manipulate the global queue lock, but in
addition it means that all threads have to check the global queue
before going to sleep (hence take a lock again) to figure how long
to sleep, and that they would all sleep only for the shortest amount
of time to the next check, one would pick it and all other ones would
go down to sleep waiting for the next check.
That's perfectly visible in time-to-first-byte measurements. A quick
test consisting in retrieving the stats page in CSV over a 48-thread
process checking 200 servers every 2 seconds shows the following tail:
percentile ttfb(ms)
99.98 2.43
99.985 5.72
99.99 32.96
99.995 82.176
99.996 82.944
99.9965 83.328
99.997 83.84
99.9975 84.288
99.998 85.12
99.9985 86.592
99.999 88
99.9995 89.728
99.9999 100.352
One solution could consist in forcefully binding checks to threads at
boot time, but that's annoying, will cause trouble for dynamic servers
and may cause some skew in the load depending on some server patterns.
Instead here we take a different approach. A check remains bound to its
thread for as long as possible, but upon every wakeup, the thread's load
is compared with another random thread's load. If it's found that that
other thread's load is less than half of the current one's, the task is
bounced to that thread. In order to prevent that new thread from doing
the same, we set a flag "CHK_ST_SLEEPING" that indicates that it just
woke up and we're bouncing the task only on this condition.
Tests have shown that the initial load was very unfair before, with a few
checks threads having a load of 15-20 and the vast majority having zero.
With this modification, after two "inter" delays, the load is either zero
or one everywhere when checks start. The same test shows a CPU usage that
significantly drops, between 0.5 and 1%. The same latency tail measurement
is much better, roughly 10 times smaller:
percentile ttfb(ms)
99.98 1.647
99.985 1.773
99.99 4.912
99.995 8.76
99.996 8.88
99.9965 8.944
99.997 9.016
99.9975 9.104
99.998 9.224
99.9985 9.416
99.999 9.8
99.9995 10.04
99.9999 10.432
In fact one difference here is that many threads work while in the past
they were waking up and going down to sleep after having perturbated the
shared lock. Thus it is anticipated that this will scale way smoother
than before. Under strace it's clearly visible that all threads are
sleeping for the time it takes to relaunch a check, there's no more
thundering herd wakeups.
However it is also possible that in some rare cases such as very short
check intervals smaller than a scheduler's timeslice (such as 4ms),
some users might have benefited from the work being concentrated on
less threads and would instead observe a small increase of apparent
CPU usage due to more total threads waking up even if that's for less
work each and less total work. That's visible with 200 servers at 4ms
where show activity shows that a few threads were overloaded and others
doing nothing. It's not a problem, though as in practice checks are not
supposed to eat much CPU and to wake up fast enough to represent a
significant load anyway, and the main issue they could have been
causing (aside the global lock) is an increase last-percentile latency.
The check struct had a "cs" field renamed to "sc", which also required
a tiny update to a few functions using it to distinguish a check from
a stream (log.c, payload.c, ssl_sample.c, tcp_sample.c, tcpcheck.c,
connection.c).
Function arguments and local variables called "cs" were renamed to "sc".
The presence of one "cs=" in the debugging traces was also turned to
"sc=" for consistency.
This renames the "struct conn_stream" to "struct stconn" and updates
the descriptions in all comments (and the rare help descriptions) to
"stream connector" or "connector". This touches a lot of files but
the change is minimal. The local variables were not even renamed, so
there's still a lot of "cs" everywhere.
Instead of using the health-check to subscribe to read/write events, we now
rely on the conn-stream. Indeed, on the server side, the conn-stream's
endpoint is a multiplexer. Thus it seems appropriate to handle subscriptions
for read/write events the same way than for the streams. Of course, the I/O
callback function is not the same. We use srv_chk_io_cb() instead of
cs_conn_io_cb().
Use include file <sys/time.h> to fix compilation error with timeval in
some files. This is as reported as 'man 7 system_data_types'. The build
error is reported on NetBSD 9.2.
This should be backported up to 2.2.
No need to include the full tree management code, type files only
need the definitions. Doing so reduces the whole code size by around
3.6% and the build time is down to just 6s.
Implement a mechanism to free a started check on runtime for dynamic
servers. A new function check_purge is created for this. The check task
will be marked for deletion and scheduled to properly close connection
elements and free the task/tasklet/buf_wait elements.
This function will be useful to delete a dynamic server wich checks.
When a tcpcheck ruleset uses multiple connections, the existing one must be
closed and destroyed before openning the new one. This part is handled in
the tcpcheck_main() function, when called from the wake callback function
(wake_srv_chk). But it is indeed a problem, because this function may be
called from the mux layer. This means a mux may call the wake callback
function of the data layer, which may release the connection and the mux. It
is easy to see how it is hazardous. And actually, depending on the
scheduling, it leads to crashes.
Thus, we must avoid to release the connection in the wake callback context,
and move this part in the check's process function instead. To do so, we
rely on the CHK_ST_CLOSE_CONN flags. When a connection must be replaced by a
new one, this flag is set on the check, in tcpcheck_main() function, and the
check's task is woken up. Then, the connection is really closed in
process_chk_conn() function.
This patch must be backported as far as 2.2, with some adaptations however
because the code is not exactly the same.
Historically, the input and output buffers of a check are allocated by hand
during the startup, with a specific size (not necessarily the same than
other buffers). But since the recent refactoring of the checks to rely
exclusively on the tcp-checks and to use the underlying mux layer, this part
is totally buggy. Indeed, because these buffers are now passed to a mux,
they maybe be swapped if a zero-copy is possible. In fact, for now it is
only possible in h2_rcv_buf(). Thus the bug concretely only exists if a h2
health-check is performed. But, it is a latent bug for other muxes.
Another problem is the size of these buffers. because it may differ for the
other buffer size, it might be source of bugs.
Finally, for configurations with hundreds of thousands of servers, having 2
buffers per check always allocated may be an issue.
To fix the bug, we now allocate these buffers when required using the buffer
pool. Thus not-running checks don't waste memory and muxes may swap them if
possible. The only drawback is the check buffers have now always the same
size than buffers used by the streams. This deprecates indirectly the
"tune.chksize" global option.
In addition, the http-check regtest have been update to perform some h2
health-checks.
Many thanks to @VigneshSP94 for its help on this bug.
This patch should solve the issue #936. It relies on the commit "MINOR:
tcpcheck: Don't handle anymore in-progress send rules in tcpcheck_main".
Both must be backport as far as 2.2.
bla
There are list definitions everywhere in the code, let's drop the need
for including list-t.h to declare them. The rest of the list manipulation
is huge however and not needed everywhere so using the list walking macros
still requires to include list.h.
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.
Checks.c remains one of the largest file of the project and it contains
too many things. The tcpchecks code represents half of this file, and
both parts are relatively isolated, so let's move it away into its own
file. We now have tcpcheck.c, tcpcheck{,-t}.h.
Doing so required to export quite a number of functions because check.c
has almost everything made static, which really doesn't help to split!
All includes that were not absolutely necessary were removed because
checks.h happens to very often be part of dependency loops. A warning
was added about this in check-t.h. The fields, enums and structs were
a bit tidied because it's particularly tedious to find anything there.
It would make sense to split this in two or more files (at least
extract tcp-checks).
The file was renamed to the singular because it was one of the rare
exceptions to have an "s" appended to its name compared to the struct
name.
