1 .. This work is licensed under a Creative Commons Attribution 4.0 International License.
2 .. http://creativecommons.org/licenses/by/4.0
3 .. (c) OPNFV, Intel Corporation and others.
5 DPDK Keep Alive Overview
6 =========================
7 SFQM aims to enable fault detection within DPDK, the very first feature to
8 meet this goal is the DPDK Keep Alive Sample app that is part of DPDK 2.2.
10 DPDK Keep Alive or KA is a sample application that acts as a heartbeat/watchdog
11 for DPDK packet processing cores, to detect application thread failure. The
12 application supports the detection of ‘failed’ DPDK cores and notification to a
13 HA/SA middleware. The purpose is to detect Packet Processing Core fails (e.g.
14 infinite loop) and ensure the failure of the core does not result in a fault
15 that is not detectable by a management entity.
17 .. Figure:: dpdk_ka.png
19 DPDK Keep Alive Sample Application
21 Essentially the app demonstrates how to detect 'silent outages' on DPDK packet
22 processing cores. The application can be decomposed into two specific parts:
23 detection and notification.
25 * The detection period is programmable/configurable but defaults to 5ms if no
27 * The Notification support is enabled by simply having a hook function that where this
28 can be 'call back support' for a fault management application with a compliant
31 DPDK Keep Alive Sample App Internals
32 ====================================
33 This section provides some explanation of the The Keep-Alive/'Liveliness'
34 conceptual scheme as well as the DPDK Keep Alive App. The initialization and
35 run-time paths are very similar to those of the L2 forwarding application (see
36 `L2 Forwarding Sample Application (in Real and Virtualized Environments)`_ for more
39 There are two types of cores: a Keep Alive Monitor Agent Core (master DPDK core)
40 and Worker cores (Tx/Rx/Forwarding cores). The Keep Alive Monitor Agent Core
41 will supervise worker cores and report any failure (2 successive missed pings).
42 The Keep-Alive/'Liveliness' conceptual scheme is:
44 * DPDK worker cores mark their liveliness as they forward traffic.
45 * A Keep Alive Monitor Agent Core runs a function every N Milliseconds to
46 inspect worker core liveliness.
47 * If keep-alive agent detects time-outs, it notifies the fault management
48 entity through a call-back function.
50 **Note:** Only the worker cores state is monitored. There is no mechanism or agent
51 to monitor the Keep Alive Monitor Agent Core.
53 DPDK Keep Alive Sample App Code Internals
54 =========================================
55 The following section provides some explanation of the code aspects that are
56 specific to the Keep Alive sample application.
58 The heartbeat functionality is initialized with a struct rte_heartbeat and the
59 callback function to invoke in the case of a timeout.
63 rte_global_keepalive_info = rte_keepalive_create(&dead_core, NULL);
64 if (rte_global_hbeat_info == NULL)
65 rte_exit(EXIT_FAILURE, "keepalive_create() failed");
67 The function that issues the pings hbeat_dispatch_pings() is configured to run
68 every check_period milliseconds.
72 if (rte_timer_reset(&hb_timer,
73 (check_period * rte_get_timer_hz()) / 1000,
76 &hbeat_dispatch_pings, rte_global_keepalive_info
78 rte_exit(EXIT_FAILURE, "Keepalive setup failure.\n");
80 The rest of the initialization and run-time path follows the same paths as the
81 the L2 forwarding application. The only addition to the main processing loop is
82 the mark alive functionality and the example random failures.
86 rte_keepalive_mark_alive(&rte_global_hbeat_info);
87 cur_tsc = rte_rdtsc();
89 /* Die randomly within 7 secs for demo purposes.. */
90 if (cur_tsc - tsc_initial > tsc_lifetime)
93 The rte_keepalive_mark_alive() function simply sets the core state to alive.
98 rte_keepalive_mark_alive(struct rte_heartbeat *keepcfg)
100 keepcfg->state_flags[rte_lcore_id()] = 1;
103 Keep Alive Monitor Agent Core Monitoring Options
104 The application can run on either a host or a guest. As such there are a number
105 of options for monitoring the Keep Alive Monitor Agent Core through a Local
106 Agent on the compute node:
108 ====================== ========== =============
109 Application Location DPDK KA LOCAL AGENT
110 ====================== ========== =============
113 ====================== ========== =============
116 For the first implementation of a Local Agent SFQM will enable:
118 ====================== ========== =============
119 Application Location DPDK KA LOCAL AGENT
120 ====================== ========== =============
122 ====================== ========== =============
124 Through extending the dpdkstat plugin for collectd with KA functionality, and
125 integrating the extended plugin with Monasca for high performing, resilient,
126 and scalable fault detection.
128 .. _L2 Forwarding Sample Application (in Real and Virtualized Environments): http://dpdk.org/doc/guides/sample_app_ug/l2_forward_real_virtual.html