From: Vallari Agrawal <val.agl002@gmail.com>
Date: Tue, 1 Oct 2024 06:10:23 +0000 (+0530)
Subject: mon: add warning NVMEOF_GATEWAY_DOWN
X-Git-Tag: v20.0.0~665^2~1
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=0006599c9e1a6941e0a347f2c86e795f1fda56c7;p=ceph.git

mon: add warning NVMEOF_GATEWAY_DOWN

In src/mon/NVMeofGwMap.cc,
add warning NVMEOF_GATEWAY_DOWN when any
gateway is in GW_UNAVAILABLE state.

Signed-off-by: Vallari Agrawal <val.agl002@gmail.com>
---

diff --git a/doc/rados/operations/health-checks.rst b/doc/rados/operations/health-checks.rst
index bbce958bebe7..4ee02c04e1eb 100644
--- a/doc/rados/operations/health-checks.rst
+++ b/doc/rados/operations/health-checks.rst
@@ -1636,8 +1636,8 @@ bucket are the same.
 NVMeoF Gateway
 --------------
 
-NVMOEF_SINGLE_GATEWAY
-__________________________________
+NVMEOF_SINGLE_GATEWAY
+_____________________
 
 One of the gateway group has only one gateway. This is not ideal because it makes
 high availability (HA) impossible with a single gatway in a group. This can lead to 
@@ -1645,6 +1645,13 @@ problems with failover and failback operations for the NVMeoF gateway.
 
 It's recommended to have multiple NVMeoF gateways in a group.
 
+NVMEOF_GATEWAY_DOWN
+___________________
+
+Some of the gateways are in the GW_UNAVAILABLE state. If a NVMeoF daemon has crashed, 
+the daemon log file (found at ``/var/log/ceph/``) may contain troubleshooting information.
+
+
 Miscellaneous
 -------------
 
diff --git a/src/mon/NVMeofGwMap.cc b/src/mon/NVMeofGwMap.cc
index c350622b7db6..b29606db998e 100755
--- a/src/mon/NVMeofGwMap.cc
+++ b/src/mon/NVMeofGwMap.cc
@@ -883,7 +883,8 @@ struct CMonRequestProposal : public Context {
 
 void NVMeofGwMap::get_health_checks(health_check_map_t *checks) const 
 {
-  list<string> detail;
+  list<string> singleGatewayDetail;
+  list<string> gatewayDownDetail;
   for (const auto& created_map_pair: created_gws) {
     const auto& group_key = created_map_pair.first;
     auto& group = group_key.second;
@@ -891,16 +892,33 @@ void NVMeofGwMap::get_health_checks(health_check_map_t *checks) const
     if ( gw_created_map.size() == 1) {
       ostringstream ss;
       ss << "NVMeoF Gateway Group '" << group << "' has 1 gateway." ;
-      detail.push_back(ss.str());
+      singleGatewayDetail.push_back(ss.str());
+    }
+    for (const auto& gw_created_pair: gw_created_map) {
+      const auto& gw_id = gw_created_pair.first;
+      const auto& gw_created  = gw_created_pair.second;
+      if (gw_created.availability == gw_availability_t::GW_UNAVAILABLE) {
+        ostringstream ss;
+        ss << "NVMeoF Gateway '" << gw_id << "' is unavailable." ;
+        gatewayDownDetail.push_back(ss.str());
+      }
     }
   }
-  if (!detail.empty()) {
+  if (!singleGatewayDetail.empty()) {
     ostringstream ss;
-    ss << detail.size() << " group(s) have only 1 nvmeof gateway"
+    ss << singleGatewayDetail.size() << " group(s) have only 1 nvmeof gateway"
       << "; HA is not possible with single gateway.";
     auto& d = checks->add("NVMEOF_SINGLE_GATEWAY", HEALTH_WARN,
-        ss.str(), detail.size());
-    d.detail.swap(detail);
+        ss.str(), singleGatewayDetail.size());
+    d.detail.swap(singleGatewayDetail);
+  }
+  if (!gatewayDownDetail.empty()) {
+    ostringstream ss;
+    ss << gatewayDownDetail.size() << " gateway(s) are in unavailable state"
+      << "; gateway might be down, try to redeploy.";
+    auto& d = checks->add("NVMEOF_GATEWAY_DOWN", HEALTH_WARN,
+        ss.str(), gatewayDownDetail.size());
+    d.detail.swap(gatewayDownDetail);
   }
 }