[ovs-dev] [PATCH] vtep: add source node replication support.
Bruce Davie
bdavie at vmware.com
Fri Apr 15 19:40:42 UTC 2016
Darrell,
Thanks for taking this on, we’ve known for a while that this should have been cleaned up.
I have a few comments and questions.
First, who decides what mode a switch is in? In most, if not all fields, either the VTEP or the controller is expected to write the value. In this case, some controllers only support service node replication, so I would expect the controller to be setting this field. However, if a controller asks a switch to enter source replication mode, and that mode is not supported by the hardware, there would need to be a way for the controller to find out that its instructions to the VTEP have been ignored.
My suggestion: make it clear in vtep.xml that this field is written by the controller (the NVC). Additionally, add a new error code by which a VTEP can announce “replication mode not supported”.
Second, I wonder if it is worth the extra complexity to have per-logical switch setting of this flag as well as per-physical switch setting. Bear in mind that a logical switch might span several physical switches, and at least in theory some of them might not support all replication modes. Seems like asking for trouble to do it this way. Also, leaving this as a per-physical switch setting only means that you won’t need to talk about the logical switch referring back to the physical switch (which has a number of problems, not least is the fact that no single physical switch “contains” the logical switch).
My suggestion: delete the sentence “A contained logical switch…physical switch.” (As noted above, logical switches may not actually be contained by a physical switch), and also delete the new flag entirely from the logical switch table.
Third, we should be clearer about the meaning of the locator_set entries in mcast_macs_remote table. Currently, when describing the locator_set entry we say:
The physical locator set to be used to reach this MAC address. In
this table, the physical locator set will be either a service node IP
address or a set of tunnel IP addresses of hypervisors (and
potentially other VTEPs).
I believe this should say:
The physical locator set to be used to reach this MAC address. In
this table, the physical locator set will be either a set of service node IP
addresses when service node replication is used or a set of tunnel IP addresses of hypervisors (and
potentially other VTEPs) when source node replication is used. In the case of service node replication,
the VTEP should send packets to one of the services nodes that is known to be reachable based on
the BFD status of the appropriate tunnel.
Finally, I would suggest some wordsmithing on the intro paragraph in the Physical Switch table:
For handling broadcast, multicast (in default manner) and unknown
unicast traffic, packets must be sent to all members of a logical switch. There are different modes
to replicate the packets. One mode of replication is to send the
traffic to a service node (a hypervisor, server or appliance) designated to handle such replication and
allow that node to replicate the traffic to other hypervisors and VTEPs.
This mode is called service node replication. A second mode of
replication, called source node replication involves the source node
sending to all hypervisors and VTEPs in which the logical switch may have ports. Hypervisors are always responsible for
doing their own replication for locally attached VMs in both modes.
(Just to be clear, there may be cases where several VTEPs have ports connected to a logical switch, and so we need to remember VTEPs when doing replication, not just hypervisors).
At this point I’ve only reviewed vtep.ovsschema and vtep.xml. Hope this makes sense & helps.
Bruce
> On Apr 14, 2016, at 2:44 PM, Darrell Ball <dlu998 at gmail.com> wrote:
>
> This patch series updates the vtep schema, vtep-ctl commands and vtep
> simulator to support source node replication in addition to service node
> replication per logical switch and per physical switch. The default
> replication mode is service_node as that was the only mode previously
> supported.
>
> Signed-off-by: Darrell Ball <dlu998 at gmail.com>
> ---
> tests/vtep-ctl.at | 16 +++++++++++++
> vtep/README.ovs-vtep.md | 18 +++++++++++++--
> vtep/ovs-vtep | 55 +++++++++++++++++++++++++++++++++++++++++----
> vtep/vtep-ctl.8.in | 20 +++++++++++++++++
> vtep/vtep-ctl.c | 60 +++++++++++++++++++++++++++++++++++++++++++++++++
> vtep/vtep.ovsschema | 13 ++++++++++-
> vtep/vtep.xml | 53 +++++++++++++++++++++++++++++++++++++++++--
> 7 files changed, 226 insertions(+), 9 deletions(-)
>
> diff --git a/tests/vtep-ctl.at b/tests/vtep-ctl.at
> index 99e97e8..8abfa8a 100644
> --- a/tests/vtep-ctl.at
> +++ b/tests/vtep-ctl.at
> @@ -264,6 +264,14 @@ CHECK_PORTS([b], [p1])
> VTEP_CTL_CLEANUP
> AT_CLEANUP
>
> +AT_SETUP([add-ps a, set-ps-replication-mode a source_node])
> +AT_KEYWORDS([vtep-ctl])
> +VTEP_CTL_SETUP
> +AT_CHECK([RUN_VTEP_CTL(
> + [add-ps a],[set-ps-replication-mode a source_node])],
> + [0], [], [], [VTEP_CTL_CLEANUP])
> +VTEP_CTL_CLEANUP
> +AT_CLEANUP
>
> dnl ----------------------------------------------------------------------
> AT_BANNER([vtep-ctl unit tests -- logical switch tests])
> @@ -318,6 +326,14 @@ CHECK_LSWITCHES([a])
> VTEP_CTL_CLEANUP
> AT_CLEANUP
>
> +AT_SETUP([add-ls a, set-ls-replication-mode a source_node])
> +AT_KEYWORDS([vtep-ctl])
> +VTEP_CTL_SETUP
> +AT_CHECK([RUN_VTEP_CTL(
> + [add-ls a],[set-ls-replication-mode a source_node])],
> + [0], [], [], [VTEP_CTL_CLEANUP])
> +VTEP_CTL_CLEANUP
> +AT_CLEANUP
>
> dnl ----------------------------------------------------------------------
> AT_BANNER([vtep-ctl unit tests -- logical binding tests])
> diff --git a/vtep/README.ovs-vtep.md b/vtep/README.ovs-vtep.md
> index 6734dab..61c937a 100644
> --- a/vtep/README.ovs-vtep.md
> +++ b/vtep/README.ovs-vtep.md
> @@ -166,13 +166,27 @@ vtep-ctl bind-ls br0 p0 0 ls0
> vtep-ctl set Logical_Switch ls0 tunnel_key=33
> ```
>
> -3. Direct unknown destinations out a tunnel:
> +3. Optionally, change the replication mode from a default of service_node to
> + source_node, which can either be done at the physical switch or logical
> + switch levels. A physical switch mode level change affects all the
> + contained logical switches in the VTEP emulator. When a logical switch is
> + first bound to a physical switch the logical switch will inherit the
> + replication mode of the physical switch. The replication mode of the
> + logical switch can be subsequently changed:
> +
> + ```
> +vtep-ctl set-ps-replication-mode br0 source_node
> + or
> +vtep-ctl set-ls-replication-mode ls0 source_node
> + ```
> +
> +4. Direct unknown destinations out a tunnel:
>
> ```
> vtep-ctl add-mcast-remote ls0 unknown-dst 10.2.2.2
> ```
>
> -4. Direct unicast destinations out a different tunnel:
> +5. Direct unicast destinations out a different tunnel:
>
> ```
> vtep-ctl add-ucast-remote ls0 00:11:22:33:44:55 10.2.2.3
> diff --git a/vtep/ovs-vtep b/vtep/ovs-vtep
> index 31ff159..da3a65d 100755
> --- a/vtep/ovs-vtep
> +++ b/vtep/ovs-vtep
> @@ -43,6 +43,8 @@ exiting = False
>
> ps_name = ""
> ps_type = ""
> +ps_replication_mode = "service_node"
> +ps_replication_mode_chg = False
> Tunnel_Ip = ""
> Lswitches = {}
> Bindings = {}
> @@ -94,6 +96,7 @@ class Logical_Switch(object):
> self.unknown_dsts = set()
> self.tunnel_key = 0
> self.setup_ls()
> + self.replication_mode = "service_node"
>
> def __del__(self):
> vlog.info("destroying lswitch %s" % self.name)
> @@ -141,13 +144,17 @@ class Logical_Switch(object):
> ovs_ofctl("add-flow %s table=1,priority=1,in_port=%s,action=%s"
> % (self.short_name, port_no, ",".join(flood_ports)))
>
> - # Traffic coming from a VTEP physical port should only be flooded to
> - # one 'unknown-dst' and to all other physical ports that belong to that
> - # VTEP device and this logical switch.
> + # Traffic coming from a VTEP physical port should always be flooded to
> + # to all the other physical ports that belong to that VTEP device and
> + # this logical switch. If the replication mode is service node then
> + # send to one unknown_dst HV (the first one here); else we assume the
> + # replication mode is source node and we send the packet to all
> + # unknown_dst HVs.
> for tunnel in self.unknown_dsts:
> port_no = self.tunnels[tunnel][0]
> flood_ports.append(port_no)
> - break
> + if self.replication_mode == "service_node":
> + break
>
> ovs_ofctl("add-flow %s table=1,priority=0,action=%s"
> % (self.short_name, ",".join(flood_ports)))
> @@ -293,8 +300,32 @@ class Logical_Switch(object):
>
> self.remote_macs = remote_macs
>
> + replication_mode = vtep_ctl("get logical_switch %s replication_mode"
> + % self.name)
> +
> + # If the logical switch level replication mode has changed then
> + # update to that value. Otherwise if the physical switch parent
> + # replication node has changed, then update the logical switch
> + # mode to the physical switch mode value.
> + replic_mode_change = False
> + if replication_mode != self.replication_mode:
> + self.replication_mode = replication_mode
> + vlog.info("%s replication mode changed to %s" %
> + (self.name, self.replication_mode))
> + replic_mode_change = True
> + elif ps_replication_mode_chg is True:
> + self.replication_mode = ps_replication_mode
> + vlog.info("%s replication mode changed to %s; inherit physical "
> + "switch replication mode"
> + % (self.name, self.replication_mode))
> + replic_mode_change = True
> +
> + unk_dsts_mode_change = False
> if (self.unknown_dsts != unknown_dsts):
> self.unknown_dsts = unknown_dsts
> + unk_dsts_mode_change = True
> +
> + if replic_mode_change is True or unk_dsts_mode_change is True:
> self.update_flood()
>
> def update_stats(self):
> @@ -575,6 +606,22 @@ def handle_physical():
> vlog.info("deleting %s from %s" % (pp_name, ps_name))
> vtep_ctl("del-port %s %s" % (ps_name, pp_name))
>
> + # Get the physical switch replication mode from the database and update
> + # the VTEP simulator physical switch replication mode and set the
> + # physical switch replication mode changed flag so the logical switches
> + # can know to update to the new physical switch mode value.
> + replication_mode = vtep_ctl("get physical_switch %s replication_mode"
> + % ps_name)
> +
> + global ps_replication_mode_chg
> + global ps_replication_mode
> + ps_replication_mode_chg = False
> + if replication_mode != ps_replication_mode:
> + ps_replication_mode = replication_mode
> + ps_replication_mode_chg = True
> + vlog.info("%s replication mode changed to %s" %
> + (ps_name, ps_replication_mode))
> +
> new_bindings = set()
> for pp_name in vtep_pp_set:
> binding_set = set(vtep_ctl("list-bindings %s %s"
> diff --git a/vtep/vtep-ctl.8.in b/vtep/vtep-ctl.8.in
> index 129c7ed..6eaf721 100644
> --- a/vtep/vtep-ctl.8.in
> +++ b/vtep/vtep-ctl.8.in
> @@ -124,6 +124,16 @@ Without \fB\-\-if\-exists\fR, attempting to delete a switch that does
> not exist is an error. With \fB\-\-if\-exists\fR, attempting to
> delete a switch that does not exist has no effect.
> .
> +.IP "\fBset\-ps\-replication\-mode \fIpswitch replication\-mode\fR"
> +Set physical switch \fIpswitch\fR replication mode to \fIreplication\-mode\fR;
> +valid replication mode values are "service_node" and "source_node".
> +The physical switch \fIpswitch\fR replication mode is inherited by all the
> +contained logical switches unless the logical switch later overrides that
> +value. When a logical switch is first bound to a physical switch
> +\fIpswitch\fR via bind\-ls, the logical switch inherits the physical switch
> +\fIpswitch\fR replication mode. This can later be overridden at the
> +logical switch level.
> +.
> .IP "\fBlist\-ps\fR"
> Lists all existing physical switches on standard output, one per line.
> .
> @@ -195,6 +205,16 @@ combination on the physical switch \fIpswitch\fR.
> List the logical switch bindings for \fIport\fR on the physical switch
> \fIpswitch\fR.
> .
> +.IP "\fBset\-ls\-replication\-mode \fIlswitch replication\-mode\fR"
> +Set logical switch \fIlswitch\fR replication mode to \fIreplication\-mode\fR;
> +valid values for replication mode are "service_node" and "source_node".
> +The physical switch replication mode is inherited by all the contained
> +logical switch \fIlswitch\fR unless the logical switch \fIlswitch\fR later
> +overrides that value. When a logical switch \fIlswitch\fR is first bound to
> +a physical switch via bind\-ls, the logical switch \fIlswitch\fR inherits the
> +physical switch replication mode. This can later be overridden at the
> +logical switch \fIlswitch\fR level.
> +.
> .SS "Logical Router Commands"
> These commands examine and manipulate logical routers.
> .
> diff --git a/vtep/vtep-ctl.c b/vtep/vtep-ctl.c
> index 29d9a17..17d57f7 100644
> --- a/vtep/vtep-ctl.c
> +++ b/vtep/vtep-ctl.c
> @@ -319,6 +319,7 @@ Manager commands:\n\
> Physical Switch commands:\n\
> add-ps PS create a new physical switch named PS\n\
> del-ps PS delete PS and all of its ports\n\
> + set-ps-replication-mode PS MODE set replication mode on PS\n\
> list-ps print the names of all the physical switches\n\
> ps-exists PS exit 2 if PS does not exist\n\
> \n\
> @@ -335,6 +336,7 @@ Logical Switch commands:\n\
> bind-ls PS PORT VLAN LS bind LS to VLAN on PORT\n\
> unbind-ls PS PORT VLAN unbind logical switch on VLAN from PORT\n\
> list-bindings PS PORT list bindings for PORT on PS\n\
> + set-ls-replication-mode LS MODE set replication mode on LS\n\
> \n\
> Logical Router commands:\n\
> add-lr LR create a new logical router named LR\n\
> @@ -846,11 +848,13 @@ pre_get_info(struct ctl_context *ctx)
> ovsdb_idl_add_column(ctx->idl, &vteprec_physical_switch_col_name);
> ovsdb_idl_add_column(ctx->idl, &vteprec_physical_switch_col_ports);
> ovsdb_idl_add_column(ctx->idl, &vteprec_physical_switch_col_tunnels);
> + ovsdb_idl_add_column(ctx->idl, &vteprec_physical_switch_col_replication_mode);
>
> ovsdb_idl_add_column(ctx->idl, &vteprec_physical_port_col_name);
> ovsdb_idl_add_column(ctx->idl, &vteprec_physical_port_col_vlan_bindings);
>
> ovsdb_idl_add_column(ctx->idl, &vteprec_logical_switch_col_name);
> + ovsdb_idl_add_column(ctx->idl, &vteprec_logical_switch_col_replication_mode);
>
> ovsdb_idl_add_column(ctx->idl, &vteprec_logical_router_col_name);
>
> @@ -1206,6 +1210,7 @@ cmd_add_ps(struct ctl_context *ctx)
>
> ps = vteprec_physical_switch_insert(ctx->txn);
> vteprec_physical_switch_set_name(ps, ps_name);
> + vteprec_physical_switch_set_replication_mode(ps, "service_node");
>
> vtep_insert_pswitch(vtepctl_ctx->vtep_global, ps);
>
> @@ -1246,6 +1251,25 @@ cmd_del_ps(struct ctl_context *ctx)
> }
>
> static void
> +cmd_set_ps_replication_mode(struct ctl_context *ctx)
> +{
> + struct vtep_ctl_context *vtepctl_ctx = vtep_ctl_context_cast(ctx);
> + struct vtep_ctl_pswitch *ps;
> +
> + vtep_ctl_context_populate_cache(ctx);
> +
> + if (strcmp(ctx->argv[2], "service_node") &&
> + strcmp(ctx->argv[2], "source_node")) {
> + ctl_fatal("Replication mode must be 'service node' or 'source_node'");
> + }
> +
> + ps = find_pswitch(vtepctl_ctx, ctx->argv[1], true);
> + vteprec_physical_switch_set_replication_mode(ps->ps_cfg, ctx->argv[2]);
> +
> + vtep_ctl_context_invalidate_cache(ctx);
> +}
> +
> +static void
> output_sorted(struct svec *svec, struct ds *output)
> {
> const char *name;
> @@ -1407,6 +1431,7 @@ cmd_add_ls(struct ctl_context *ctx)
>
> ls = vteprec_logical_switch_insert(ctx->txn);
> vteprec_logical_switch_set_name(ls, ls_name);
> + vteprec_logical_switch_set_replication_mode(ls, "service_node");
>
> vtep_ctl_context_invalidate_cache(ctx);
> }
> @@ -1492,6 +1517,8 @@ cmd_bind_ls(struct ctl_context *ctx)
> struct vtep_ctl_lswitch *ls;
> struct vtep_ctl_port *port;
> const char *vlan;
> + struct vtep_ctl_pswitch *ps;
> + const char *ps_name = ctx->argv[1];
>
> vtep_ctl_context_populate_cache(ctx);
>
> @@ -1502,6 +1529,15 @@ cmd_bind_ls(struct ctl_context *ctx)
> add_ls_binding_to_cache(port, vlan, ls);
> commit_ls_bindings(port);
>
> + ps = find_pswitch(vtepctl_ctx, ps_name, true);
> +
> + if (!ps->ps_cfg->replication_mode) {
> + ctl_fatal("Replication mode not found for physical switch %s",
> + ps_name);
> + }
> + vteprec_logical_switch_set_replication_mode(ls->ls_cfg,
> + ps->ps_cfg->replication_mode);
> +
> vtep_ctl_context_invalidate_cache(ctx);
> }
>
> @@ -1523,6 +1559,26 @@ cmd_unbind_ls(struct ctl_context *ctx)
> vtep_ctl_context_invalidate_cache(ctx);
> }
>
> +static void
> +cmd_set_ls_replication_mode(struct ctl_context *ctx)
> +{
> + struct vtep_ctl_context *vtepctl_ctx = vtep_ctl_context_cast(ctx);
> + struct vtep_ctl_lswitch *ls;
> + const char *ls_name = ctx->argv[1];
> +
> + vtep_ctl_context_populate_cache(ctx);
> +
> + if (strcmp(ctx->argv[2], "service_node") &&
> + strcmp(ctx->argv[2], "source_node")) {
> + ctl_fatal("Replication mode must be 'service node' or 'source_node'");
> + }
> +
> + ls = find_lswitch(vtepctl_ctx, ls_name, true);
> + vteprec_logical_switch_set_replication_mode(ls->ls_cfg, ctx->argv[2]);
> +
> + vtep_ctl_context_invalidate_cache(ctx);
> +}
> +
> static struct vtep_ctl_lrouter *
> find_lrouter(struct vtep_ctl_context *vtepctl_ctx,
> const char *name, bool must_exist)
> @@ -2443,6 +2499,8 @@ static const struct ctl_command_syntax vtep_commands[] = {
> {"del-ps", 1, 1, NULL, pre_get_info, cmd_del_ps, NULL, "--if-exists", RW},
> {"list-ps", 0, 0, NULL, pre_get_info, cmd_list_ps, NULL, "", RO},
> {"ps-exists", 1, 1, NULL, pre_get_info, cmd_ps_exists, NULL, "", RO},
> + {"set-ps-replication-mode", 2, 2, "PS MODE", pre_get_info,
> + cmd_set_ps_replication_mode, NULL, "", RW},
>
> /* Port commands. */
> {"list-ports", 1, 1, NULL, pre_get_info, cmd_list_ports, NULL, "", RO},
> @@ -2459,6 +2517,8 @@ static const struct ctl_command_syntax vtep_commands[] = {
> {"list-bindings", 2, 2, NULL, pre_get_info, cmd_list_bindings, NULL, "", RO},
> {"bind-ls", 4, 4, NULL, pre_get_info, cmd_bind_ls, NULL, "", RO},
> {"unbind-ls", 3, 3, NULL, pre_get_info, cmd_unbind_ls, NULL, "", RO},
> + {"set-ls-replication-mode", 2, 2, "LS MODE", pre_get_info,
> + cmd_set_ls_replication_mode, NULL, "", RW},
>
> /* Logical Router commands. */
> {"add-lr", 1, 1, NULL, pre_get_info, cmd_add_lr, NULL, "--may-exist", RW},
> diff --git a/vtep/vtep.ovsschema b/vtep/vtep.ovsschema
> index 533fd2e..abb83b8 100644
> --- a/vtep/vtep.ovsschema
> +++ b/vtep/vtep.ovsschema
> @@ -1,6 +1,7 @@
> {
> "name": "hardware_vtep",
> - "cksum": "770244945 11113",
> + "version": "1.1.0",
> + "cksum": "213455912 11473",
> "tables": {
> "Global": {
> "columns": {
> @@ -28,6 +29,11 @@
> "tunnels": {
> "type": {"key": {"type": "uuid", "refTable": "Tunnel"},
> "min": 0, "max": "unlimited"}},
> + "replication_mode": {
> + "type": {
> + "key": {
> + "enum": ["set", ["source_node","service_node"]],
> + "type": "string"}}},
> "other_config": {
> "type": {"key": "string", "value": "string",
> "min": 0, "max": "unlimited"}},
> @@ -96,6 +102,11 @@
> "name": {"type": "string"},
> "description": {"type": "string"},
> "tunnel_key": {"type": {"key": "integer", "min": 0, "max": 1}},
> + "replication_mode": {
> + "type": {
> + "key": {
> + "enum": ["set", ["source_node","service_node"]],
> + "type": "string"}}},
> "other_config": {
> "type": {"key": "string", "value": "string",
> "min": 0, "max": "unlimited"}}},
> diff --git a/vtep/vtep.xml b/vtep/vtep.xml
> index a3a6988..a70c188 100644
> --- a/vtep/vtep.xml
> +++ b/vtep/vtep.xml
> @@ -304,6 +304,31 @@
> </column>
> </group>
>
> + <group title="Per Physical-Switch Replication Mode">
> + <p>
> + For handling broadcast, multicast (in default manner) and unknown
> + unicast traffic, packets can be sent to all members of a logical
> + switch contained by a physical switch. There are different modes
> + to replicate the packets. One mode of replication is to send the
> + traffic to a hypervisor designated to handle such replication and
> + allow that hypervisor to replicate the traffic to other hypervisors.
> + This mode is called service node replication. A second mode of
> + replication, called source node replication involves the source node
> + sending to all hypervisors. Hypervisors are always responsible for
> + doing their own replication for locally attached VMs in both modes.
> + </p>
> +
> + <column name="replication_mode">
> + <p>
> + This column defines the replication mode per
> + <ref table="Physical_Switch"/>. There are two valid values
> + presently, <code>service_node</code> and <code>source_node</code>.
> + A contained logical switch may reference the value set by the
> + parent physical switch.
> + </p>
> + </column>
> + </group>
> +
> <group title="Identification">
> <column name="name">
> Symbolic name for the switch, such as its hostname.
> @@ -754,6 +779,30 @@
> </column>
> </group>
>
> + <group title="Per Logical-Switch Replication Mode">
> + <p>
> + For handling broadcast, unknown unicast and multicast traffic
> + (in default manner), packets can be sent to all members of a logical
> + switch. There are different modes to replicate the packets. One
> + mode of replication is to send the traffic to a hypervisor designated
> + to handle such replication and allow that hypervisor to replicate the
> + traffic to other hypervisors. This mode is called service node
> + replication. A second mode of replication, called source node
> + replication involves the source node sending to all hypervisors.
> + Hypervisors are always responsible for doing their own replication
> + for locally attached VMs in both modes.
> + </p>
> +
> + <column name="replication_mode">
> + <p>
> + This column defines the replication mode per
> + <ref table="Logical_Switch"/>. There are two valid values presently,
> + <code>service_node</code> and <code>source_node</code>.
> + </p>
> +
> + </column>
> + </group>
> +
> <group title="Identification">
> <column name="name">
> Symbolic name for the logical switch.
> @@ -887,8 +936,8 @@
> Multicast packet replication may be handled by a service node,
> in which case the physical locators will be IP addresses of
> service nodes. If the VTEP supports replication onto multiple
> - tunnels, then this may be used to replicate directly onto
> - VTEP-hypervisor tunnels.
> + tunnels, using source node replication, then this may be used to
> + replicate directly onto VTEP-hypervisor tunnels.
> </p>
>
> <column name="MAC">
> --
> 1.9.1
>
More information about the dev
mailing list