Skip to contentSkip to conversation

ipvs: Add simple weighted failover scheduler

PATCH In review Kenny Mathis 3 participants 6 messages Sep 9, 2014 – Sep 10, 2014 1 message not in the archive

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 ipvs: Add simple weighted failover scheduler

Please consider this patch for addition in 3.17

Add simple weighted IPVS failover support to the Linux kernel. All other scheduling modules implement some form of load balancing, while this offers a simple failover solution. Connections are directed to the appropriate server based solely on highest weight value and server availability. Tested functionality with keepalived.

---
 net/netfilter/ipvs/Kconfig    |   10 ++++++
 net/netfilter/ipvs/Makefile   |    1 +
 net/netfilter/ipvs/ip_vs_fo.c |   79 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 90 insertions(+)
 create mode 100644 net/netfilter/ipvs/ip_vs_fo.c
net/netfilter/ipvs/Kconfig +10 -0
index 0c3b167..3b6929d 100644
@@ -152,6 +152,16 @@ config	IP_VS_WLC
 	  If you want to compile it in kernel, say Y. To compile it as a
 	  module, choose M here. If unsure, say N.
 
+config  IP_VS_FO
+		tristate "weighted failover scheduling"
+	---help---
+	  The weighted failover scheduling algorithm directs network
+	  connections to the server with the highest weight that is
+	  currently available.
+
+	  If you want to compile it in kernel, say Y. To compile it as a
+	  module, choose M here. If unsure, say N.
+
 config	IP_VS_LBLC
 	tristate "locality-based least-connection scheduling"
 	---help---
net/netfilter/ipvs/Makefile +1 -0
index 34ee602..38b2723 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_IP_VS_RR) += ip_vs_rr.o
 obj-$(CONFIG_IP_VS_WRR) += ip_vs_wrr.o
 obj-$(CONFIG_IP_VS_LC) += ip_vs_lc.o
 obj-$(CONFIG_IP_VS_WLC) += ip_vs_wlc.o
+obj-$(CONFIG_IP_VS_FO) += ip_vs_fo.o
 obj-$(CONFIG_IP_VS_LBLC) += ip_vs_lblc.o
 obj-$(CONFIG_IP_VS_LBLCR) += ip_vs_lblcr.o
 obj-$(CONFIG_IP_VS_DH) += ip_vs_dh.o
net/netfilter/ipvs/ip_vs_fo.c new file +79 -0

79 lines not on this page · raw message

Signature · 1 line
1.7.9.5
maintainer Acked-by on v2 1/1
#1 lore raw

Hello,

Quoted 9 lines · Kenny Mathis

Please consider this patch for addition in 3.17

Add simple weighted IPVS failover support to the Linux kernel. All other scheduling modules implement some form of load balancing, while this offers a simple failover solution. Connections are directed to the appropriate server based solely on highest weight value and server availability. Tested functionality with keepalived.

Looks good to me, thanks!

Simon, you can apply it to ipvs-next if there are
no any objections.

Quoted 127 lines
---
 net/netfilter/ipvs/Kconfig    |   10 ++++++
 net/netfilter/ipvs/Makefile   |    1 +
 net/netfilter/ipvs/ip_vs_fo.c |   79 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 90 insertions(+)
 create mode 100644 net/netfilter/ipvs/ip_vs_fo.c
net/netfilter/ipvs/Kconfig +10 -0
index 0c3b167..3b6929d 100644
@@ -152,6 +152,16 @@ config	IP_VS_WLC
 	  If you want to compile it in kernel, say Y. To compile it as a
 	  module, choose M here. If unsure, say N.
 
+config  IP_VS_FO
+		tristate "weighted failover scheduling"
+	---help---
+	  The weighted failover scheduling algorithm directs network
+	  connections to the server with the highest weight that is
+	  currently available.
+
+	  If you want to compile it in kernel, say Y. To compile it as a
+	  module, choose M here. If unsure, say N.
+
 config	IP_VS_LBLC
 	tristate "locality-based least-connection scheduling"
 	---help---
net/netfilter/ipvs/Makefile +1 -0
index 34ee602..38b2723 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_IP_VS_RR) += ip_vs_rr.o
 obj-$(CONFIG_IP_VS_WRR) += ip_vs_wrr.o
 obj-$(CONFIG_IP_VS_LC) += ip_vs_lc.o
 obj-$(CONFIG_IP_VS_WLC) += ip_vs_wlc.o
+obj-$(CONFIG_IP_VS_FO) += ip_vs_fo.o
 obj-$(CONFIG_IP_VS_LBLC) += ip_vs_lblc.o
 obj-$(CONFIG_IP_VS_LBLCR) += ip_vs_lblcr.o
 obj-$(CONFIG_IP_VS_DH) += ip_vs_dh.o
net/netfilter/ipvs/ip_vs_fo.c new file +79 -0
new file mode 100644
index 0000000..6a2647d
@@ -0,0 +1,79 @@
+/*
+ * IPVS:        Weighted Fail Over module
+ *
+ * Authors:     Kenny Mathis <[email protected]>
+ *
+ *              This program is free software; you can redistribute it and/or
+ *              modify it under the terms of the GNU General Public License
+ *              as published by the Free Software Foundation; either version
+ *              2 of the License, or (at your option) any later version.
+ *
+ * Changes:
+ *     Kenny Mathis            :     added initial functionality based on weight
+ *
+ */
+
+#define KMSG_COMPONENT "IPVS"
+#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+
+#include <net/ip_vs.h>
+
+/* Weighted Fail Over Module */
+static struct ip_vs_dest *
+ip_vs_fo_schedule(struct ip_vs_service *svc, const struct sk_buff *skb,
+		  struct ip_vs_iphdr *iph)
+{
+	struct ip_vs_dest *dest, *hweight = NULL;
+	int hw = 0; /* Track highest weight */
+
+	IP_VS_DBG(6, "ip_vs_fo_schedule(): Scheduling...\n");
+
+	/* Basic failover functionality
+	 * Find virtual server with highest weight and send it traffic
+	 */
+	list_for_each_entry_rcu(dest, &svc->destinations, n_list) {
+		if (!(dest->flags & IP_VS_DEST_F_OVERLOAD) &&
+		    atomic_read(&dest->weight) > hw) {
+			hweight = dest;
+			hw = atomic_read(&dest->weight);
+		}
+	}
+
+	if (hweight) {
+		IP_VS_DBG_BUF(6, "FO: server %s:%u activeconns %d weight %d\n",
+			      IP_VS_DBG_ADDR(svc->af, &hweight->addr),
+			      ntohs(hweight->port),
+			      atomic_read(&hweight->activeconns),
+			      atomic_read(&hweight->weight));
+		return hweight;
+	}
+
+	ip_vs_scheduler_err(svc, "no destination available");
+	return NULL;
+}
+
+static struct ip_vs_scheduler ip_vs_fo_scheduler = {
+	.name =			"fo",
+	.refcnt =		ATOMIC_INIT(0),
+	.module =		THIS_MODULE,
+	.n_list =		LIST_HEAD_INIT(ip_vs_fo_scheduler.n_list),
+	.schedule =		ip_vs_fo_schedule,
+};
+
+static int __init ip_vs_fo_init(void)
+{
+	return register_ip_vs_scheduler(&ip_vs_fo_scheduler);
+}
+
+static void __exit ip_vs_fo_cleanup(void)
+{
+	unregister_ip_vs_scheduler(&ip_vs_fo_scheduler);
+	synchronize_rcu();
+}
+
+module_init(ip_vs_fo_init);
+module_exit(ip_vs_fo_cleanup);
+MODULE_LICENSE("GPL");
Signature · 1 line
1.7.9.5

Regards

--
Julian Anastasov <[email protected]>

maintainer
#2 lore raw
Quoted 6 lines · Julian Anastasov

Hello,

Quoted 1 line · Kenny Mathis

Please consider this patch for addition in 3.17

There is no need to update this patch, but for reference
the line above belongs below scissors ('---') below your Signed-off-by line. i.e. just above the diffstat below.

That way git will automatically remove it when you apply your patch.

Quoted 6 lines
Quoted 6 lines

Add simple weighted IPVS failover support to the Linux kernel. All other scheduling modules implement some form of load balancing, while this offers a simple failover solution. Connections are directed to the appropriate server based solely on highest weight value and server availability. Tested functionality with keepalived.

Quoted 8 lines
Quoted 1 line

Looks good to me, thanks!

Simon, you can apply it to ipvs-next if there are
no any objections.

I have no objections. I have applied it.

Quoted 133 lines
Quoted 127 lines
---
 net/netfilter/ipvs/Kconfig    |   10 ++++++
 net/netfilter/ipvs/Makefile   |    1 +
 net/netfilter/ipvs/ip_vs_fo.c |   79 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 90 insertions(+)
 create mode 100644 net/netfilter/ipvs/ip_vs_fo.c
net/netfilter/ipvs/Kconfig +10 -0
index 0c3b167..3b6929d 100644
@@ -152,6 +152,16 @@ config	IP_VS_WLC
 	  If you want to compile it in kernel, say Y. To compile it as a
 	  module, choose M here. If unsure, say N.
 
+config  IP_VS_FO
+		tristate "weighted failover scheduling"
+	---help---
+	  The weighted failover scheduling algorithm directs network
+	  connections to the server with the highest weight that is
+	  currently available.
+
+	  If you want to compile it in kernel, say Y. To compile it as a
+	  module, choose M here. If unsure, say N.
+
 config	IP_VS_LBLC
 	tristate "locality-based least-connection scheduling"
 	---help---
net/netfilter/ipvs/Makefile +1 -0
index 34ee602..38b2723 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_IP_VS_RR) += ip_vs_rr.o
 obj-$(CONFIG_IP_VS_WRR) += ip_vs_wrr.o
 obj-$(CONFIG_IP_VS_LC) += ip_vs_lc.o
 obj-$(CONFIG_IP_VS_WLC) += ip_vs_wlc.o
+obj-$(CONFIG_IP_VS_FO) += ip_vs_fo.o
 obj-$(CONFIG_IP_VS_LBLC) += ip_vs_lblc.o
 obj-$(CONFIG_IP_VS_LBLCR) += ip_vs_lblcr.o
 obj-$(CONFIG_IP_VS_DH) += ip_vs_dh.o
net/netfilter/ipvs/ip_vs_fo.c new file +79 -0
new file mode 100644
index 0000000..6a2647d
@@ -0,0 +1,79 @@
+/*
+ * IPVS:        Weighted Fail Over module
+ *
+ * Authors:     Kenny Mathis <[email protected]>
+ *
+ *              This program is free software; you can redistribute it and/or
+ *              modify it under the terms of the GNU General Public License
+ *              as published by the Free Software Foundation; either version
+ *              2 of the License, or (at your option) any later version.
+ *
+ * Changes:
+ *     Kenny Mathis            :     added initial functionality based on weight
+ *
+ */
+
+#define KMSG_COMPONENT "IPVS"
+#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+
+#include <net/ip_vs.h>
+
+/* Weighted Fail Over Module */
+static struct ip_vs_dest *
+ip_vs_fo_schedule(struct ip_vs_service *svc, const struct sk_buff *skb,
+		  struct ip_vs_iphdr *iph)
+{
+	struct ip_vs_dest *dest, *hweight = NULL;
+	int hw = 0; /* Track highest weight */
+
+	IP_VS_DBG(6, "ip_vs_fo_schedule(): Scheduling...\n");
+
+	/* Basic failover functionality
+	 * Find virtual server with highest weight and send it traffic
+	 */
+	list_for_each_entry_rcu(dest, &svc->destinations, n_list) {
+		if (!(dest->flags & IP_VS_DEST_F_OVERLOAD) &&
+		    atomic_read(&dest->weight) > hw) {
+			hweight = dest;
+			hw = atomic_read(&dest->weight);
+		}
+	}
+
+	if (hweight) {
+		IP_VS_DBG_BUF(6, "FO: server %s:%u activeconns %d weight %d\n",
+			      IP_VS_DBG_ADDR(svc->af, &hweight->addr),
+			      ntohs(hweight->port),
+			      atomic_read(&hweight->activeconns),
+			      atomic_read(&hweight->weight));
+		return hweight;
+	}
+
+	ip_vs_scheduler_err(svc, "no destination available");
+	return NULL;
+}
+
+static struct ip_vs_scheduler ip_vs_fo_scheduler = {
+	.name =			"fo",
+	.refcnt =		ATOMIC_INIT(0),
+	.module =		THIS_MODULE,
+	.n_list =		LIST_HEAD_INIT(ip_vs_fo_scheduler.n_list),
+	.schedule =		ip_vs_fo_schedule,
+};
+
+static int __init ip_vs_fo_init(void)
+{
+	return register_ip_vs_scheduler(&ip_vs_fo_scheduler);
+}
+
+static void __exit ip_vs_fo_cleanup(void)
+{
+	unregister_ip_vs_scheduler(&ip_vs_fo_scheduler);
+	synchronize_rcu();
+}
+
+module_init(ip_vs_fo_init);
+module_exit(ip_vs_fo_cleanup);
+MODULE_LICENSE("GPL");
Signature · 1 line
1.7.9.5

Regards

--
Julian Anastasov <[email protected]>

maintainer
#3 lore raw
Quoted 22 lines · Simon Horman
Quoted 6 lines · Julian Anastasov

Hello,

Quoted 1 line · Kenny Mathis

Please consider this patch for addition in 3.17

There is no need to update this patch, but for reference
the line above belongs below scissors ('---') below your Signed-off-by line. i.e. just above the diffstat below.

That way git will automatically remove it when you apply your patch.

Quoted 8 lines
Quoted 7 lines

Add simple weighted IPVS failover support to the Linux kernel. All other scheduling modules implement some form of load balancing, while this offers a simple failover solution. Connections are directed to the appropriate server based solely on highest weight value and server availability. Tested functionality with keepalived.

Quoted 8 lines
Quoted 6 lines

Looks good to me, thanks!

Simon, you can apply it to ipvs-next if there are
no any objections.

I have no objections. I have applied it.

Kenny, are you planning an update to ipvsadm to allow it to use the new scheduler?

author
#4 lore raw
Quoted 1 line

From what I can tell, ipvsadm doesn't need any specific code updates.

The only place that the schedulers are actually listed is in the man page and the SCHEDULERS file, but I can submit a change to ensure it's listed. When I tested with keepalived, the failover (fo) module showed up properly under ipvsadm as well without any modifications.

Quoted 34 lines · Simon Horman
Quoted 22 lines · Simon Horman
Quoted 6 lines · Julian Anastasov

Hello,

Quoted 1 line · Kenny Mathis

Please consider this patch for addition in 3.17

There is no need to update this patch, but for reference
the line above belongs below scissors ('---') below your Signed-off-by line. i.e. just above the diffstat below.

That way git will automatically remove it when you apply your patch.

Quoted 8 lines
Quoted 7 lines

Add simple weighted IPVS failover support to the Linux kernel. All other scheduling modules implement some form of load balancing, while this offers a simple failover solution. Connections are directed to the appropriate server based solely on highest weight value and server availability. Tested functionality with keepalived.

Quoted 8 lines
Quoted 6 lines

Looks good to me, thanks!

Simon, you can apply it to ipvs-next if there are
no any objections.

I have no objections. I have applied it.

Kenny, are you planning an update to ipvsadm to allow it to use the new scheduler?

A message that is not in the archive (sent off-list or not yet received): CAPvRj6herHa82QA9iszdT1+AMu05UJML427W4wct14NZ=suYZA@mail.gmail.com

maintainer
#6 lore raw
Quoted 5 lines · Kenny Mathis

From what I can tell, ipvsadm doesn't need any specific code updates. The only place the schedulers are actually listed is in the man page and the SCHEDULERS file, so I could submit a change to ensure it's listed. When I tested with keepalived, the failover (fo) module showed up properly under ipvsadm as well without any modifications.

Thanks, I understand.

I think updating SCHEDULERS would be good but clearly its not urgent.