glusterfs.git/rpc/rpc-lib/src/rpc-clnt.h, branch release-3.11

Halo Replication feature for AFR translator

2017-05-08T05:37:07+00:00

	Backport of https://review.gluster.org/16177
		    https://review.gluster.org/17174

Merged both these patches to make sure IPV6 changes don't make it to 3.11 at all.

Summary:
Halo Geo-replication is a feature which allows Gluster or NFS clients to write
locally to their region (as defined by a latency "halo" or threshold if you
like), and have their writes asynchronously propagate from their origin to the
rest of the cluster.  Clients can also write synchronously to the cluster
simply by specifying a halo-latency which is very large (e.g. 10seconds) which
will include all bricks.

In other words, it allows clients to decide at mount time if they desire
synchronous or asynchronous IO into a cluster and the cluster can support both
of these modes to any number of clients simultaneously.

There are a few new volume options due to this feature:
  halo-shd-latency:  The threshold below which self-heal daemons will
  consider children (bricks) connected.

  halo-nfsd-latency: The threshold below which NFS daemons will consider
  children (bricks) connected.

  halo-latency: The threshold below which all other clients will
  consider children (bricks) connected.

  halo-min-replicas: The minimum number of replicas which are to
  be enforced regardless of latency specified in the above 3 options.
  If the number of children falls below this threshold the next
  best (chosen by latency) shall be swapped in.

New FUSE mount options:
  halo-latency & halo-min-replicas: As descripted above.

This feature combined with multi-threaded SHD support (D1271745) results in
some pretty cool geo-replication possibilities.

Operational Notes:
- Global consistency is gaurenteed for synchronous clients, this is provided by
  the existing entry-locking mechanism.
- Asynchronous clients on the other hand and merely consistent to their region.
  Writes & deletes will be protected via entry-locks as usual preventing
  concurrent writes into files which are undergoing replication.  Read operations
  on the other hand should never block.
- Writes are allowed from _any_ region and propagated from the origin to all
  other regions.  The take away from this is care should be taken to ensure
  multiple writers do not write the same files resulting in a gfid split-brain
  which will require resolution via split-brain policies (majority, mtime &
  size).  Recommended method for preventing this is using the nfs-auth feature to
  define which region for each share has RW permissions, tiers not in the origin
  region should have RO perms.

TODO:
- Synchronous clients (including the SHD) should choose clients from their own
  region as preferred sources for reads.  Most of the plumbing is in place for
  this via the child_latency array.
- Better GFID split brain handling & better dent type split brain handling
  (i.e. create a trash can and move the offending files into it).
- Tagging in addition to latency as a means of defining which children you wish
  to synchronously write to

Test Plan:
- The usual suspects, clang, gcc w/ address sanitizer & valgrind
- Prove tests

Reviewers: jackl, dph, cjh, meyering

Reviewed By: meyering

Subscribers: ethanr

Differential Revision: https://phabricator.fb.com/D1272053

Tasks: 4117827

 >Change-Id: I694a9ab429722da538da171ec528406e77b5e6d1
 >BUG: 1428061
 >Signed-off-by: Kevin Vigor 
 >Reviewed-on: http://review.gluster.org/16099
 >Reviewed-on: https://review.gluster.org/16177
 >Tested-by: Pranith Kumar Karampuri 
 >Smoke: Gluster Build System 
 >NetBSD-regression: NetBSD Build System 
 >CentOS-regression: Gluster Build System 
 >Reviewed-by: Pranith Kumar Karampuri 

BUG: 1448416
Change-Id: I694a9ab429722da538da171ec528406e77b5e6d1
Signed-off-by: Pranith Kumar K 
Reviewed-on: https://review.gluster.org/17192
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Kaushal M

rpc/clnt: remove locks while notifying CONNECT/DISCONNECT

2017-03-01T14:35:48+00:00

Locking during notify was introduced as part of commit
aa22f24f5db7659387704998ae01520708869873 [1]. The fix was introduced
to fix out-of-order CONNECT/DISCONNECT events from rpc-clnt to parent
xlators [2]. However as part of handling DISCONNECT protocol/client
does unwind saved frames (with failure) waiting for responses. This
saved_frames_unwind can be a costly operation and hence ideally
shouldn't be included in the critical section of notifylock, as it
unnecessarily delays the reconnection to same brick. Also, its not a
good practise to pass control to other xlators holding a lock as it
can lead to deadlocks. So, this patch removes locking in rpc-clnt
while notifying parent xlators.

To fix [2], two changes are present in this patch:

* notify DISCONNECT before cleaning up rpc connection (same as commit
  a6b63e11b7758cf1bfcb6798, patch [3]).
* protocol/client uses rpc_clnt_cleanup_and_start, which cleans up rpc
  connection and does a start while handling a DISCONNECT event from
  rpc. Note that patch [3] was reverted as rpc_clnt_start called in
  quick_reconnect path of protocol/client didn't invoke connect on
  transport as the connection was not cleaned up _yet_ (as cleanup was
  moved post notification in rpc-clnt). This resulted in clients never
  attempting connect to bricks.

Note that one of the neater ways to fix [2] (without using locks) is
to introduce generation numbers to map CONNECT and DISCONNECTS across
epochs and ignore DISCONNECT events if they don't belong to current
epoch. However, this approach is a bit complex to implement and
requires time. So, current patch is a hacky stop-gap fix till we come
up with a more cleaner solution.

[1] http://review.gluster.org/15916
[2] https://bugzilla.redhat.com/show_bug.cgi?id=1386626
[3] http://review.gluster.org/15681

Change-Id: I62daeee8bb1430004e28558f6eb133efd4ccf418
Signed-off-by: Raghavendra G 
BUG: 1427012
Reviewed-on: https://review.gluster.org/16784
Smoke: Gluster Build System 
Reviewed-by: Milind Changire 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System

core: run many bricks within one glusterfsd process

2017-01-31T00:13:58+00:00

This patch adds support for multiple brick translator stacks running
in a single brick server process.  This reduces our per-brick memory usage by
approximately 3x, and our appetite for TCP ports even more.  It also creates
potential to avoid process/thread thrashing, and to improve QoS by scheduling
more carefully across the bricks, but realizing that potential will require
further work.

Multiplexing is controlled by the "cluster.brick-multiplex" global option.  By
default it's off, and bricks are started in separate processes as before.  If
multiplexing is enabled, then *compatible* bricks (mostly those with the same
transport options) will be started in the same process.

Change-Id: I45059454e51d6f4cbb29a4953359c09a408695cb
BUG: 1385758
Signed-off-by: Jeff Darcy 
Reviewed-on: https://review.gluster.org/14763
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Vijay Bellur

rpc: fix for race between rpc and protocol/client

2016-12-05T14:09:48+00:00

It is possible that the notification thread which notifies
protocol/client layer about the disconnection is put to sleep
and meanwhile, a fuse thread or a timer thread initiates and
completes reconnection to the brick. The notification thread
is then woken up and protocol/client layer updates its flags
to indicate that network is disconnected. No reconnection is
initiated because reconnection is rpc-lib layer's responsibility
and its flags indicate that connection is connected.

Fix: Serialize connect and disconnect notify

Credit: Raghavendra Talur 
Change-Id: I8ff5d1a3283b47f5c26848a42016a40bc34ffc1d
BUG: 1386626
Signed-off-by: Rajesh Joseph 
Reviewed-on: http://review.gluster.org/15916
Reviewed-by: Raghavendra G 
Smoke: Gluster Build System 
Reviewed-by: Raghavendra Talur 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System

changelog/rpc: Fix rpc_clnt_t mem leaks

2016-07-22T15:12:52+00:00

PROBLEM:
   1. Freeing up rpc_clnt object might lead to crashes. Well,
      it was not a necessity to free rpc-clnt object till now
      because all the existing use cases needs to reconnect
      back on disconnects. Hence timer code was not taking
      ref on rpc-clnt object.

      Glusterd had some use-cases that led to crash due to
      ping-timer and they fixed only those code paths that
      involve ping-timer.

      Now, since changelog has an use-case where rpc-clnt
      need to be freed up, we need to fix timer code to take
      refs

   2. In changelog, because of issue 1, only mydata was being
      freed which is incorrect. And there are races where
      rpc-clnt object would access the freed mydata which
      would lead to crashes.

      Since changelog xlator resides on brick side and is long
      living process, if multiple libgfchangelog consumers
      register to changelog and disconnect/reconnect mulitple
      times, it would result in leak of 'rpc-clnt' object
      for every connect/disconnect.

SOLUTION:
   1. Handle ref/unref of 'rpc_clnt' structure in timer
      functions properly.
   2. In changelog, unref 'rpc_clnt' in RPC_CLNT_DISCONNECT
      after disabling timers and free mydata on RPC_CLNT_DESTROY.

RPC SETUP IN CHANGELOG:
   1. changelog xlator initiates rpc server say 'changelog_rpc_server'
   2. libgfchangelog initiates one rpc server say 'libgfchangelog_rpc_server'
   3. libgfchangelog initiates rpc client and connects to 'changelog_rpc_server'
   4. In return changelog_rpc_server initiates a rpc client and connects back
      to 'libgfchangelog_rpc_server'

REF/UNREF HANDLING IN TIMER FUNCTIONS:
Let's say rpc clnt refcount = 1
   1. Take the ref before reigstering callback to timer queue
           >>>>  rpc_clnt_ref (say ref count becomes = 2)
   2. Register a callback to timer say 'callback1'
   3. If register fails:
           >>>> rpc_clnt_unref (ref count = 1)
   4. On timer expiration, 'callback1' gets called. So unref rpc clnt at the end
      in 'callback1'. This is corresponding to ref taken in step 1
           >>>> rpc_clnt_unref (ref count = 1)
   5. The cycle from step-1 to step-4 continues....until timer cancel event happens
   6. timer cancel of say 'callback1'
           If timer cancel fails:
                 Do nothing, Step-4 would have unrefd
           If timer cancel succeeds:
                 >>>> rpc_clnt_unref (ref count = 1)

Change-Id: I91389bc511b8b1a17824941970ee8d2c29a74a09
BUG: 1316178
Signed-off-by: Kotresh HR 
Reviewed-on: http://review.gluster.org/13658
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Raghavendra G

glusterd/rpc : Discard duplicate Disconnect events

2016-03-22T19:24:53+00:00

If a peer rpc disconnect event has been already processed, skip the furthers as
processing them are overheads and sometimes may lead to a crash like due to a
double free

Change-Id: Iec589ce85daf28fd5b267cb6fc82a4238e0e8adc
BUG: 1318546
Signed-off-by: Atin Mukherjee 
Reviewed-on: http://review.gluster.org/13790
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Jeff Darcy

rpc: add owner xlator argument to rpc_clnt_new

2015-08-13T06:12:50+00:00

The @owner argument tells RPC layer the xlator that owns
the connection and to which xlator THIS needs be set during
network notifications like CONNECT and DISCONNECT.

Code paths that originate from the head of a (volume) graph and use
STACK_WIND ensure that the RPC local endpoint has the right xlator saved
in the frame of the call (callback pair). This guarantees that the
callback is executed in the right xlator context.

The client handshake process which includes fetching of brick ports from
glusterd, setting lk-version on the brick for the session, don't have
the correct xlator set in their frames. The problem lies with RPC
notifications. It doesn't have the provision to set THIS with the xlator
that is registered with the corresponding RPC programs. e.g,
RPC_CLNT_CONNECT event received by protocol/client doesn't have THIS set
to its xlator. This implies, call(-callbacks) originating from this
thread don't have the right xlator set too.

The fix would be to save the xlator registered with the RPC connection
during rpc_clnt_new. e.g, protocol/client's xlator would be saved with
the RPC connection that it 'owns'. RPC notifications such as CONNECT,
DISCONNECT, etc inherit THIS from the RPC connection's xlator.

Change-Id: I9dea2c35378c511d800ef58f7fa2ea5552f2c409
BUG: 1235582
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/11436
Tested-by: Gluster Build System 
Tested-by: NetBSD Build System 
Reviewed-by: Raghavendra G

client, rpc: make ping-timeout configurable for glusterfs clients

2015-08-05T07:24:56+00:00

Change-Id: Idd94adb0457aaffce7330f56f98cebafa2c4dae8
BUG: 1249499
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/11818
Tested-by: Gluster Build System 
Reviewed-by: Raghavendra G 
Tested-by: NetBSD Build System

glusterd: nfs,shd,quotad,snapd daemons refactoring

2015-02-20T12:04:08+00:00

This patch ports nfs, shd, quotad & snapd with the approach suggested in
http://www.gluster.org/pipermail/gluster-devel/2014-December/043180.html

Change-Id: I4ea5b38793f87fc85cc9d2cf873727351dedffd2
BUG: 1191486
Signed-off-by: Atin Mukherjee 
Signed-off-by:  Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/9428
Tested-by: Gluster Build System 
Reviewed-by: Anand Nekkunti

client/rpc: Add ping and msgs stats

2014-07-31T16:03:56+00:00

This can be seen as below,

># cat $META/graphs/active/vol-client-0/private |grep ping_msgs_sent
ping_msgs_sent = 2

># cat $META/graphs/active/vol-client-0/private |grep "^msgs_sent"
msgs_sent = 13

where $META is //.meta

Change-Id: I2107ec2b045bac701377760635e18758adb943a3
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/8285
Tested-by: Gluster Build System 
Reviewed-by: Atin Mukherjee 
Reviewed-by: Anand Avati