commit bdd403c58703522dcbcf402039936eeb579ac7bf
author Serge Bazanski <q3k@hackerspace.pl> Thu Oct 28 23:37:38 2021 +0000
committer q3k <q3k@hackerspace.pl> Mon Dec 13 22:30:46 2021 +0000
tree 951c8ab991b48e88269af43640ee481203e433ed
parent 8469691645b4e16f10e517cbd6b7b9850011b273

cluster: k0: move cockroachdb away from bc01n01, fixup joins

Reminded by a power failure on bc01n0{1,2}, we migrate away from at
least one of them into another server.

We also fix up the startup join parameter to not include the node itself
(which is not necessary, but a nice thing to have nonetheless).

Since bc01n01 was the initial node of the cluster, we also disable the
init job for k0 (which we don't care about anyway).

Change-Id: I3406471c0f9542e9d802d39138e400b5a5e74794
Reviewed-on: https://gerrit.hackerspace.pl/c/hscloud/+/1176
Reviewed-by: q3k <q3k@hackerspace.pl>

cluster/kube/k0.libsonnet

diff --git a/cluster/kube/k0.libsonnet b/cluster/kube/k0.libsonnet
index d2a54fc..9fea81c 100644
--- a/cluster/kube/k0.libsonnet
+++ b/cluster/kube/k0.libsonnet
@@ -64,9 +64,9 @@
             waw2: cockroachdb.Cluster("crdb-waw1") {
                 cfg+: {
                     topology: [
-                        { name: "bc01n01", node: "bc01n01.hswaw.net" },
                         { name: "bc01n02", node: "bc01n02.hswaw.net" },
                         { name: "dcr01s22", node: "dcr01s22.hswaw.net" },
+                        { name: "dcr01s24", node: "dcr01s24.hswaw.net" },
                     ],
                     // Host path on SSD.
                     hostPath: "/var/db/crdb-waw1",
@@ -74,6 +74,7 @@
                         "crdb-waw1.hswaw.net",
                     ],
                 },
+                initJob:: null,
             },
             clients: {
                 cccampix: k0.cockroach.waw2.Client("cccampix"),

cluster/kube/lib/cockroachdb.libsonnet

diff --git a/cluster/kube/lib/cockroachdb.libsonnet b/cluster/kube/lib/cockroachdb.libsonnet
index 2486570..dff2ebd 100644
--- a/cluster/kube/lib/cockroachdb.libsonnet
+++ b/cluster/kube/lib/cockroachdb.libsonnet
@@ -215,6 +215,13 @@
                         },
                     },
                     spec+: {
+                        strategy+: {
+                          type: "RollingUpdate",
+                          rollingUpdate: {
+                            maxSurge: 0,
+                            maxUnavailable: 1,
+                          },
+                        },
                         template+: {
                             metadata: server.deploy.metadata,
                             spec+: {
@@ -287,7 +294,7 @@
                                             "--certs-dir", "/cockroach/cockroach-certs",
                                             "--advertise-host", "%s.cluster.local" % server.service.host, 
                                             "--cache", "25%", "--max-sql-memory", "25%",
-                                            "--join", std.join(",", ["%s.cluster.local:%d" % [s.service.host, cluster.cfg.portServe] for s in cluster.servers]),
+                                            "--join", std.join(",", ["%s.cluster.local:%d" % [s.service.host, cluster.cfg.portServe] for s in cluster.servers if s.service.host != server.service.host]),
                                             "--listen-addr=0.0.0.0:%d" % cluster.cfg.portServe,
                                             "--http-addr=0.0.0.0:%d" % cluster.cfg.portHttp,
                                         ],