The problem is that zclient->redist[ZEBRA_ROUTE_MAX] used for storing a
client’s redist state, has no address-family qualification. This means
a client can only store its interest in a protocol (connected, static etc.),
but cant choose IPv4 or ipv6 with that. This hindered implementation on
client sides to manage redistribution of ipv4 and ipv6 both.
BGP's redistribution of protocols like connected/static is one such place.
One fix could be to overload this and flap the redist connection each time
any new afi is added for redist, but that may have side-effects on the
existing afi redist.
The cleaner way is to modify redist data-structure to also take AFI, and adjust
routines that deal with it, so that a client can register for a protocol
redistribution based on the AFI. BGP already maintains redistribution state
based on afi and protocol (bgp->redist[AFI_MAX][ZEBRA_ROUTE_MAX]). This patch
takes care of filling up the gap in zclient/zserv redistribution state to
also use AFI qualification.
Signed-off-by: Vipin Kumar <vipin@cumulusnetworks.com>
Reviewed-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Dinesh G Dutt <ddutt@cumulusnetworks.com>
——————————————-------------
- etc/init.d/quagga is modified to support creating separate ospf daemon
process for each instance. Each individual instance is monitored by
watchquagga just like any protocol daemons.(requires initd-mi.patch).
- Vtysh is modified to able to connect to multiple daemons of the same
protocol (supported for OSPF only for now).
- ospfd is modified to remember the Instance-ID that its invoked with. For
the entire life of the process it caters to any command request that
matches that instance-ID (unless its a non instance specific command).
Routes/messages to zebra are tagged with instance-ID.
- zebra route/redistribute mechanisms are modified to work with
[protocol type + instance-id]
- bgpd now has ability to have multiple instance specific redistribution
for a protocol (OSPF only supported/tested for now).
- zlog ability to display instance-id besides the protocol/daemon name.
- Changes in other daemons are to because of the needed integration with
some of the modified APIs/routines. (Didn’t prefer replicating too many
separate instance specific APIs.)
- config/show/debug commands are modified to take instance-id argument
as appropriate.
Guidelines to start using multi-instance ospf
---------------------------------------------
The patch is backward compatible, i.e for any previous way of single ospf
deamon(router ospf <cr>) will continue to work as is, including all the
show commands etc.
To enable multiple instances, do the following:
1. service quagga stop
2. Modify /etc/quagga/daemons to add instance-ids of each desired
instance in the following format:
ospfd=“yes"
ospfd_instances="1,2,3"
assuming you want to enable 3 instances with those instance ids.
3. Create corresponding ospfd config files as ospfd-1.conf, ospfd-2.conf
and ospfd-3.conf.
4. service quagga start/restart
5. Verify that the deamons are started as expected. You should see
ospfd started with -n <instance-id> option.
ps –ef | grep quagga
With that /var/run/quagga/ should have ospfd-<instance-id>.pid and
ospfd-<instance-id>/vty to each instance.
6. vtysh to work with instances as you would with any other deamons.
7. Overall most quagga semantics are the same working with the instance
deamon, like it is for any other daemon.
NOTE:
To safeguard against errors leading to too many processes getting invoked,
a hard limit on number of instance-ids is in place, currently its 5.
Allowed instance-id range is <1-65535>
Once daemons are up, show running from vtysh should show the instance-id
of each daemon as 'router ospf <instance-id>’ (without needing explicit
configuration)
Instance-id can not be changed via vtysh, other router ospf configuration
is allowed as before.
Signed-off-by: Vipin Kumar <vipin@cumulusnetworks.com>
Reviewed-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Dinesh G Dutt <ddutt@cumulusnetworks.com>
Reapply the bits of revision b4e45f6 that pertain to isisd. That is:
git log -n 1 -p b4e45f6 -- isisd
These were dropped in the merge with the Google ISIS code (14d2bbaa).
The commit message for b4e45f6 is:
fix zebra protocol after MP-BGP changes
The previous commits modified both zebra and bgpd for additional
SAFI field, but not any other routing daemon, which led to zebra
daemon crashing with failed assertion.
2006-01-17 Paul Jakma <paul.jakma@sun.com>
* bgp_nexthop.c: (zlookup_read*) convert to new Zserv format.
(zlookup_query_ipv6) ditto.
(bgp_import_check) ditto.
* isis_zebra.c: (isis_zebra_route_add_ipv4) fix for new
zserv format.
ISIS_ROUTE_FLAG_ACTIVE flag before running SPF.
* isisd.[ch]: Separate route tables for different levels. SPF is done
separately, but in case of L1L2 area they have to be merged.
* isis_zebra.c: Set/unset ISIS_ROUTE_FLAG_ZEBRA_SYNC flag correctly in
case of adding/removing IPv4 routes.
* zebra_route.c: Rework route validating process. Merging L1 and L2
tables in case of L1L2 area.
In short - many changes to make SPF work more correctly, add/remove
to/from RIB also works now. It's still very far from perfect though.
Implement non-blocking zclient I/O with buffering.
* zclient.h (struct zclient): Add two fields to support non-blocking
I/O: struct buffer *wb, and struct thread *t_write.
(zclient_free): Remove function.
(zebra_redistribute_send): Change 2nd arg from socket fd to
struct zclient * (needed to support non-blocking I/O and buffering).
(zclient_send_message): New function to send an arbitrary
message with non-blocking I/O.
* zclient.c (zclient_new): Create write buffer.
(zclient_free): Remove unused function.
(zclient_stop): Must cancel new t_write thread. Also, reset
all buffers: ibuf, obuf, and wb.
(zclient_failed): New helper function for typical error handling.
(zclient_flush_data): New thread to flush queued data.
(zclient_send_message): New function to send the message in
zclient->obuf to zebra using non-blocking I/O and buffering.
(zebra_message_send, zapi_ipv4_route, zapi_ipv6_route): Use
new zclient_send_message function instead of calling writen.
(zclient_start): Set socket non-blocking. Also, change 2nd arg
to zebra_redistribute_send from zclient->sock to zclient.
(zebra_redistribute_send): Change 2nd arg to struct zclient *.
Can now use zclient->obuf to assemble the message instead of
allocating a temporary stream. And call zclient_send_message to
send the message instead of writen.
(zclient_read): Convert to support non-blocking I/O by using
stream_read_try instead of deprecated stream_read.
(zclient_redistribute): Change 2nd arg to zebra_redistribute_send
from zclient->sock to zclient.
* ospf6_zebra.c (ospf6_zebra_redistribute, ospf6_zebra_no_redistribute):
Change 2nd arg to zebra_redistribute_send from zclient->sock
to zclient.
* ospf_zebra.c (ospf_zebra_add): Call zclient_send_message instead
of writen.
* rip_zebra.c (rip_redistribute_set, rip_redistribute_unset,
rip_redistribute_clean): Change 2nd arg to zebra_redistribute_send
from zclient->sock to zclient.
* ripng_zebra.c (ripng_redistribute_unset, ripng_redistribute_clean):
Change 2nd arg to zebra_redistribute_send from zclient->sock
to zclient.
* bgp_zebra.c (bgp_redistribute_set, bgp_redistribute_unset):
The 2nd arg to zebra_redistribute_send is now zclient instead of
zclient->sock.
* isis_zebra.h (isis_zebra_finish): Remove declaration of unused
function.
* isis_zebra.c (isis_zebra_route_add_ipv4): Call zclient_send_message
to send the message to zebra instead of calling writen directly, since
zclient_send_message understands non-blocking I/O and will manage
the buffer queue appropriately.
(isis_zebra_finish): Remove unused function, particularly since
the zclient_free function has been removed.
* (global): Fix up list loops to match changes in lib/linklist,
and some basic auditing of usage.
* configure.ac: define QUAGGA_NO_DEPRECATED_INTERFACES
* HACKING: Add notes about deprecating interfaces and commands.
* lib/linklist.h: Add usage comments.
Rename getdata macro to listgetdata.
Rename nextnode to listnextnode and fix its odd behaviour to be
less dangerous.
Make listgetdata macro assert node is not null, NULL list entries
should be bug condition.
ALL_LIST_ELEMENTS, new macro, forward-referencing macro for use
with for loop, Suggested by Jim Carlson of Sun.
Add ALL_LIST_ELEMENTS_RO for cases which obviously do not need the
"safety" of previous macro.
LISTNODE_ADD and DELETE macros renamed to ATTACH, DETACH, to
distinguish from the similarly named functions, and reflect their
effect better.
Add a QUAGGA_NO_DEPRECATED_INTERFACES define guarded section
with the old defines which were modified above,
for backwards compatibility - guarded to prevent Quagga using it..
* lib/linklist.c: fix up for linklist.h changes.
* ospf6d/ospf6_abr.c: (ospf6_abr_examin_brouter) change to a single
scan of the area list, rather than scanning all areas first for
INTER_ROUTER and then again for INTER_NETWORK. According to
16.2, the scan should be area specific anyway, and further
ospf6d does not seem to implement 16.3 anyway.
Fix problems when netlink interfaces are renamed (same ifindex used
for a new interface). Start cleaning up some problems with the way
interface names are handled.
* interface.c: (if_new_intern_ifindex) Remove obsolete function.
(if_delete_update) After distributing the interface deletion message,
set ifp->ifindex to IFINDEX_INTERNAL.
(if_dump_vty) Detect pseudo interface by checking if ifp->ifindex is
IFINDEX_INTERNAL.
(zebra_interface) Check return code from interface_cmd.func.
Do not set internal ifindex values to if_new_intern_ifindex(),
since we now use IFINDEX_INTERNAL for all pseudo interfaces.
* kernel_socket.c: (ifm_read) Fix code and comments to reflect that
all internal interfaces now have ifp->ifindex set to IFINDEX_INTERNAL.
* rt_netlink.c: (set_ifindex) New function used to update ifp->ifindex.
Detects interface rename events by checking if that ifindex is already
being used. If it is, delete the old interface before assigning
the ifindex to the new interface.
(netlink_interface, netlink_link_change) Call set_ifindex to update
the ifindex.
* if.h: Remove define for IFINDEX_INTERNBASE and add define
IFINDEX_INTERNAL 0, since all internal (i.e. non-kernel) pseudo-
interfaces should have ifindex set to 0.
(if_new) Remove function.
(if_delete_retain) New function to delete an interface without
removing from iflist and freeing the structure.
(ifname2ifindex) New function.
* if.c: (if_new) Remove function (absorb into if_create).
(if_create) Replace function if_new with call to calloc.
Set ifp->ifindex to IFINDEX_INTERNAL. Fix off-by-one error
in assert to check length of interface name. Add error message
if interface with this name already exists.
(if_delete_retain) New function to delete an interface without
removing from iflist and freeing the structure.
(if_delete) Implement with help of if_delete_retain.
(ifindex2ifname) Reimplement using if_lookup_by_index.
(ifname2ifindex) New function to complement ifindex2ifname.
(interface) The interface command should check the name length
and fail with a warning message if it is too long.
(no_interface) Fix spelling in warning message.
(if_nametoindex) Reimplement using if_lookup_by_name.
(if_indextoname, ifaddr_ipv4_lookup) Reimplement using
if_lookup_by_index.
* bgp_zebra.c: (bgp_interface_delete) After deleting, set ifp->ifindex
to IFINDEX_INTERNAL.
* isis_zebra.c: (isis_zebra_if_del) Call if_delete_retain instead
of if_delete, since it is generally not safe to remove interface
structures. After deleting, set ifp->ifindex to IFINDEX_INTERNAL.
(zebra_interface_if_lookup) Tighten up code.
* ospf6_zebra.c: (ospf6_zebra_if_del) Previously, this whole function
was commented out. But this is not safe: we should at least update
the ifindex when the interface is deleted. So the new version
updates the interface status and sets ifp->ifindex to
IFINDEX_INTERNAL.
(ospf6_zebra_route_update) Use if_indextoname properly.
* ospf_vty.c: (show_ip_ospf_interface_sub) Show ifindex and interface
flags to help with debugging.
* ospf_zebra.c: (ospf_interface_delete) After deleting, set ifp->ifindex
to IFINDEX_INTERNAL.
(zebra_interface_if_lookup) Make function static. Tighten up code.
* rip_interface.c: (rip_interface_delete) After deleting, set
ifp->ifindex to IFINDEX_INTERNAL.
* ripng_interface.c: (ripng_interface_delete) After deleting, set
ifp->ifindex to IFINDEX_INTERNAL.
