glusterfs.git/rpc/rpc-lib/src/rpc-transport.c, branch v8.0

rpc-transport: minor changes

2020-01-13T07:28:53+00:00

- Removed dead code
- Remove redundant checks
- Changed dict functions to use dict_..._sizen() functions.

Change-Id: If00aaa90eef4078effd5b7fed2294f872e001b0a
updates: bz#1193929
Signed-off-by: Yaniv Kaul

multiple files: another attempt to remove includes

2019-06-14T16:50:32+00:00

There are many include statements that are not needed.
A previous more ambitious attempt failed because of *BSD plafrom
(see https://review.gluster.org/#/c/glusterfs/+/21929/ )

Now trying a more conservative reduction.
It does not solve all circular deps that we have, but it
does reduce some of them. There is just too much to handle
reasonably (dht-common.h includes dht-lock.h which includes
dht-common.h ...), but it does reduce the overall number of lines
of include we need to look at in the future to understand and fix
the mess later one.

Change-Id: I550cd001bdefb8be0fe67632f783c0ef6bee3f9f
updates: bz#1193929
Signed-off-by: Yaniv Kaul

rpc/transport: Missing a ref on dict while creating transport object

2019-03-20T13:24:44+00:00

while creating rpc_tranpsort object, we store a dictionary without
taking a ref on dict but it does an unref during the cleaning of the
transport object.

So the rpc layer expect the caller to take a ref on the dictionary
before passing dict to rpc layer. This leads to a lot of confusion
across the code base and leads to ref leaks.

Semantically, this is not correct. It is the rpc layer responsibility
to take a ref when storing it, and free during the cleanup.

I'm listing down the total issues or leaks across the code base because
of this confusion. These issues are currently present in the upstream
master.

1) changelog_rpc_client_init

2) quota_enforcer_init

3) rpcsvc_create_listeners : when there are two transport, like tcp,rdma.

4) quotad_aggregator_init

5) glusterd: init

6) nfs3_init_state

7) server: init

8) client:init

This patch does the cleanup according to the semantics.

Change-Id: I46373af9630373eb375ee6de0e6f2bbe2a677425
updates: bz#1659708
Signed-off-by: Mohammed Rafi KC

core: implement a global thread pool

2019-02-18T02:58:24+00:00

This patch implements a thread pool that is wait-free for adding jobs to
the queue and uses a very small locked region to get jobs. This makes it
possible to decrease contention drastically. It's based on wfcqueue
structure provided by urcu library.

It automatically enables more threads when load demands it, and stops
them when not needed. There's a maximum number of threads that can be
used. This value can be configured.

Depending on the workload, the maximum number of threads plays an
important role. So it needs to be configured for optimal performance.
Currently the thread pool doesn't self adjust the maximum for the
workload, so this configuration needs to be changed manually.

For this reason, the global thread pool has been made optional, so that
volumes can still use the thread pool provided by io-threads.

To enable it for bricks, the following option needs to be set:

   config.global-threading = on

This option has no effect if bricks are already running. A restart is
required to activate it. It's recommended to also enable the following
option when running bricks with the global thread pool:

   performance.iot-pass-through = on

To enable it for a FUSE mount point, the option '--global-threading'
must be added to the mount command. To change it, an umount and remount
is needed. It's recommended to disable the following option when using
global threading on a mount point:

   performance.client-io-threads = off

To enable it for services managed by glusterd, glusterd needs to be
started with option '--global-threading'. In this case all daemons, like
self-heal, will be using the global thread pool.

Currently it can only be enabled for bricks, FUSE mounts and glusterd
services.

The maximum number of threads for clients and bricks can be configured
using the following options:

   config.client-threads
   config.brick-threads

These options can be applied online and its effect is immediate most of
the times. If one of them is set to 0, the maximum number of threads
will be calcutated as #cores * 2.

Some distributions use a very old userspace-rcu library (version 0.7)
for this reason, some header files from version 0.10 have been copied
into contrib/userspace-rcu and are used if the detected version is 0.7
or older.

An additional change has been made to io-threads to prevent that threads
are started when iot-pass-through is set.

Change-Id: I09d19e246b9e6d53c6247b29dfca6af6ee00a24b
updates: #532
Signed-off-by: Xavi Hernandez

core: heketi-cli is throwing error "target is busy"

2019-01-24T06:54:41+00:00

Problem: At the time of deleting block hosting volume
         through heketi-cli , it is throwing an error "target is busy".
         cli is throwing an error because brick is not detached successfully
         and brick is not detached due to race condition to cleanp xprt
         associated with detached brick

Solution: To avoid xprt specifc race condition introduce an atomic flag
          on rpc_transport

Change-Id: Id4ff1fe8375a63be71fb3343f455190a1b8bb6d4
fixes: bz#1668190
Signed-off-by: Mohit Agrawal

rpc: use address-family option from vol file

2019-01-22T13:47:19+00:00

This patch helps enable IPv6 connections in the cluster.
The default address-family is IPv4 without using this option explicitly.

When address-family is set to "inet6" in the /etc/glusterfs/glusterd.vol
file, the mount command-line also needs to have
-o xlator-option="transport.address-family=inet6" added to it.

This option also gets added to the brick command-line.
Snapshot and gfapi use-cases should also use this option to pass in the
inet6 address-family.

Change-Id: I97db91021af27bacb6d7578e33ea4817f66d7270
fixes: bz#1635863
Signed-off-by: Milind Changire

libglusterfs: Move devel headers under glusterfs directory

2018-12-05T21:47:04+00:00

libglusterfs devel package headers are referenced in code using
include semantics for a program, this while it works can be better
especially when dealing with out of tree xlator builds or in
general out of tree devel package usage.

Towards this, the following changes are done,
- moved all devel headers under a glusterfs directory
- Included these headers using system header notation <> in all
code outside of libglusterfs
- Included these headers using own program notation "" within
libglusterfs

This change although big, is just moving around the headers and
making it correct when including these headers from other sources.

This helps us correctly include libglusterfs includes without
namespace conflicts.

Change-Id: Id2a98854e671a7ee5d73be44da5ba1a74252423b
Updates: bz#1193929
Signed-off-by: ShyamsundarR

rpc: check if fini is there before calling it

2018-12-04T02:27:03+00:00

The rpc_transport_t structure is allocated and filled in the
rpc_transport_load function. If filling the fileds of the rpc
structure fails, then in the failure handling the structure is
freed by rpc_transport_cleanup. There, it unconditionally calls
fini. But, if the failure handling was invoked because of any
failure in between the allocation of rpc_transport_t and filling
the transport->fini (including the failure to fill fini ()), then
rpc_transport_cleanup can lead to a segfault.

Change-Id: I8be9b84cd6b19933c559c9736198a6e440373f68
fixes: bz#1654917
Signed-off-by: Raghavendra Bhat

server: Resolve memory leak path in server_init

2018-12-03T11:34:35+00:00

Problem: 1) server_init does not cleanup allocate resources
            while it is failed before return error
         2) dict leak at the time of graph destroying

Solution: 1) free resources in case of server_init is failed
          2) Take dict_ref of graph xlator before destroying
             the graph to avoid leak

Change-Id: I9e31e156b9ed6bebe622745a8be0e470774e3d15
fixes: bz#1654917
Signed-off-by: Mohit Agrawal

rpcsvc: provide each request handler thread its own queue

2018-11-29T01:19:12+00:00

A single global per program queue is contended by all request handler
threads and event threads. This can lead to high contention. So,
reduce the contention by providing each request handler thread its own
private queue.

Thanks to "Manoj Pillai" for the idea of pairing a
single queue with a fixed request-handler-thread and event-thread,
which brought down the performance regression due to overhead of
queuing significantly.

Thanks to "Xavi Hernandez" for discussion on
how to communicate the event-thread death to request-handler-thread.

Thanks to "Karan Sandha" for voluntarily running
the perf benchmarks to qualify that performance regression introduced
by ping-timer-fixes is fixed with this patch and patiently running
many iterations of regression tests while RCAing the issue.

Thanks to "Milind Changire" for patiently running
the many iterations of perf benchmarking tests while RCAing the
regression caused by ping-timer-expiry fixes.

Change-Id: I578c3fc67713f4234bd3abbec5d3fbba19059ea5
Fixes: bz#1644629
Signed-off-by: Raghavendra Gowdappa