Yield more information about services to client

api/admin/admin.proto

@@ -50,6 +50,10 @@ message QueryListItem {
     string Description = 2;
     string VmStatus = 3;
     string TrustLevel = 4;
+    string VmType = 5;
+    string ServiceType = 6;
+    optional string VmName = 7; // None for host running services
+    optional string AgentName = 8; // None for agents
 }
 
 message QueryListResponse {

client/src/client.rs

@@ -11,6 +11,7 @@ pub use givc_common::query::{Event, QueryResult};
 use givc_common::types::*;
 
 use crate::endpoint::{EndpointConfig, TlsConfig};
+use crate::error::rewrap_error;
 
 type Client = pb::admin_service_client::AdminServiceClient<Channel>;
 
@@ -94,7 +95,12 @@ impl AdminClient {
             vm_name,
             args,
         };
-        let _response = self.connect_to().await?.start_application(request).await?;
+        let _response = self
+            .connect_to()
+            .await?
+            .start_application(request)
+            .await
+            .map_err(rewrap_error)?;
         Ok(())
     }
 
@@ -104,7 +110,12 @@ impl AdminClient {
             vm_name: None,
             args: Vec::new(),
         };
-        let _response = self.connect_to().await?.stop_application(request).await?;
+        let _response = self
+            .connect_to()
+            .await?
+            .stop_application(request)
+            .await
+            .map_err(rewrap_error)?;
         Ok(())
     }
 
@@ -114,7 +125,12 @@ impl AdminClient {
             vm_name: None,
             args: Vec::new(),
         };
-        let _response = self.connect_to().await?.pause_application(request).await?;
+        let _response = self
+            .connect_to()
+            .await?
+            .pause_application(request)
+            .await
+            .map_err(rewrap_error)?;
         Ok(())
     }
 
@@ -124,31 +140,56 @@ impl AdminClient {
             vm_name: None,
             args: Vec::new(),
         };
-        let _response = self.connect_to().await?.resume_application(request).await?;
+        let _response = self
+            .connect_to()
+            .await?
+            .resume_application(request)
+            .await
+            .map_err(rewrap_error)?;
         Ok(())
     }
 
     pub async fn reboot(&self) -> anyhow::Result<()> {
         let request = pb::admin::Empty {};
-        let _response = self.connect_to().await?.reboot(request).await?;
+        let _response = self
+            .connect_to()
+            .await?
+            .reboot(request)
+            .await
+            .map_err(rewrap_error)?;
         Ok(())
     }
 
     pub async fn poweroff(&self) -> anyhow::Result<()> {
         let request = pb::admin::Empty {};
-        let _response = self.connect_to().await?.poweroff(request).await?;
+        let _response = self
+            .connect_to()
+            .await?
+            .poweroff(request)
+            .await
+            .map_err(rewrap_error)?;
         Ok(())
     }
 
     pub async fn suspend(&self) -> anyhow::Result<()> {
         let request = pb::admin::Empty {};
-        let _response = self.connect_to().await?.suspend(request).await?;
+        let _response = self
+            .connect_to()
+            .await?
+            .suspend(request)
+            .await
+            .map_err(rewrap_error)?;
         Ok(())
     }
 
     pub async fn wakeup(&self) -> anyhow::Result<()> {
         let request = pb::admin::Empty {};
-        let _response = self.connect_to().await?.wakeup(request).await?;
+        let _response = self
+            .connect_to()
+            .await?
+            .wakeup(request)
+            .await
+            .map_err(rewrap_error)?;
         Ok(())
     }
 
@@ -165,7 +206,8 @@ impl AdminClient {
         self.connect_to()
             .await?
             .query_list(pb::admin::Empty {})
-            .await?
+            .await
+            .map_err(rewrap_error)?
             .into_inner()
             .list
             .into_iter()
@@ -177,15 +219,17 @@ impl AdminClient {
         self.connect_to()
             .await?
             .set_locale(pb::admin::LocaleRequest { locale })
-            .await?;
+            .await
+            .map_err(rewrap_error)?;
         Ok(())
     }
 
     pub async fn set_timezone(&self, timezone: String) -> anyhow::Result<()> {
         self.connect_to()
             .await?
             .set_timezone(pb::admin::TimezoneRequest { timezone })
-            .await?;
+            .await
+            .map_err(rewrap_error)?;
         Ok(())
     }
 
@@ -199,7 +243,8 @@ impl AdminClient {
             .connect_to()
             .await?
             .watch(pb::admin::Empty {})
-            .await?
+            .await
+            .map_err(rewrap_error)?
             .into_inner();
 
         let list = match watch.try_next().await? {

client/src/endpoint.rs

@@ -95,7 +95,8 @@ impl EndpointConfig {
         let url = transport_config_to_url(&self.transport.address, self.tls.is_some());
         info!("Connecting to {url}, TLS name {:?}", &self.tls);
         let mut endpoint = Endpoint::try_from(url.clone())?
-            .timeout(Duration::from_secs(5))
+            .timeout(Duration::from_millis(500))
+            .connect_timeout(Duration::from_millis(300))
             .concurrency_limit(30);
         if let Some(tls) = &self.tls {
             endpoint = endpoint.tls_config(tls.client_config()?)?;

client/src/error.rs

@@ -0,0 +1,17 @@
+use anyhow::Error;
+use tonic::Status;
+use tonic_types::StatusExt;
+use tracing::{debug, error};
+
+pub fn rewrap_error(status: Status) -> Error {
+    let mut err = Error::msg(status.message().to_owned());
+    let details = status.get_error_details();
+    if let Some(debug_info) = details.debug_info() {
+        err = err.context(format!("Detail: {}", debug_info.detail));
+        err = debug_info
+            .stack_entries
+            .iter()
+            .fold(err, |err, each| err.context(format!("Stack: {}", each)))
+    };
+    err
+}

client/src/lib.rs

@@ -1,3 +1,4 @@
 pub mod client;
 pub mod endpoint;
+pub mod error;
 pub use crate::client::{AdminClient, QueryResult};

common/src/query.rs

@@ -1,4 +1,5 @@
 // Types related to QueryList and Watch API
+use super::types::{ServiceType, VmType};
 use crate::pb;
 use pb::admin::watch_item::Status;
 
@@ -33,6 +34,10 @@ pub struct QueryResult {
     pub description: String, //App name, some details
     pub status: VMStatus,
     pub trust_level: TrustLevel,
+    pub vm_type: VmType,
+    pub service_type: ServiceType,
+    pub vm_name: Option<String>,
+    pub agent_name: Option<String>,
 }
 
 impl QueryResult {
@@ -51,6 +56,12 @@ impl TryFrom<pb::QueryListItem> for QueryResult {
                 .with_context(|| format!("While parsing vm_status {}", item.vm_status))?,
             trust_level: TrustLevel::from_str(item.trust_level.as_str())
                 .with_context(|| format!("While parsing trust_level {}", item.trust_level))?,
+            vm_type: VmType::from_str(item.vm_type.as_str())
+                .with_context(|| format!("While parsing vm_type {}", item.vm_type))?,
+            service_type: ServiceType::from_str(item.service_type.as_str())
+                .with_context(|| format!("While parsing service_type {}", item.service_type))?,
+            agent_name: item.agent_name,
+            vm_name: item.vm_name,
         })
     }
 }
@@ -62,6 +73,10 @@ impl From<QueryResult> for pb::QueryListItem {
             description: val.description,
             vm_status: val.status.to_string(),
             trust_level: val.trust_level.to_string(),
+            vm_type: val.vm_type.to_string(),
+            service_type: val.service_type.to_string(),
+            agent_name: val.agent_name,
+            vm_name: val.vm_name,
         }
     }
 }

common/src/types.rs

@@ -5,6 +5,8 @@ use crate::pb;
 use std::convert::{Into, TryFrom};
 
 use anyhow::bail;
+use serde::Serialize;
+use strum::{Display, EnumString};
 use tokio_vsock::VsockAddr;
 
 #[derive(Debug, Copy, Clone, PartialEq)]
@@ -13,15 +15,15 @@ pub struct UnitType {
     pub service: ServiceType,
 }
 
-#[derive(Debug, Copy, Clone, PartialEq)]
+#[derive(Debug, Copy, Clone, PartialEq, Serialize, EnumString, Display)]
 pub enum VmType {
     Host,
     AdmVM,
     SysVM,
     AppVM,
 }
 
-#[derive(Debug, Copy, Clone, PartialEq)]
+#[derive(Debug, Copy, Clone, PartialEq, Serialize, EnumString, Display)]
 pub enum ServiceType {
     Mgr,
     Svc,

nixos/modules/admin.nix

@@ -137,7 +137,7 @@ in
           "HOST_KEY" = "${cfg.tls.keyPath}";
         }
         // attrsets.optionalAttrs cfg.debug {
-          "RUST_BACKTRACE" = "1";
+#          "RUST_BACKTRACE" = "1";
           "GIVC_LOG" = "debug";
         };
     };

nixos/tests/admin.nix

@@ -81,11 +81,15 @@ in
                 ];
                 tls = mkTls "ghaf-host";
               };
-              systemd.services."microvm@foot-vm" = {
+              systemd.services."microvm@chromium-vm" = {
+                preStart = ''
+                  expr $(cat /tmp/[email protected] 2>/dev/null || true) + 1 >/tmp/[email protected]
+                '';
                 script = ''
                   # Do nothing script, simulating microvm service
                   while true; do sleep 10; done
                 '';
+                wantedBy = [ "multi-user.target" ];
               };
             };
             guivm =
@@ -321,6 +325,7 @@ in
               with subtest("Clean run"):
                   print(hostvm.succeed("${cli} --addr ${nodes.adminvm.config.givc.admin.addr} --port ${nodes.adminvm.config.givc.admin.port} --cacert ${nodes.hostvm.givc.host.tls.caCertPath} --cert ${nodes.hostvm.givc.host.tls.certPath} --key ${nodes.hostvm.givc.host.tls.keyPath} ${if tls then "" else "--notls"} --name ${nodes.adminvm.config.givc.admin.name} start --vm chromium-vm foot"))
                   time.sleep(10) # Give few seconds to application to spin up
+                  print(hostvm.succeed("cat /tmp/[email protected]"))
                   wait_for_window("ghaf@appvm")
 
               with subtest("crash and restart"):
@@ -356,6 +361,15 @@ in
                   print(hostvm.succeed("${cli} --addr ${nodes.adminvm.config.givc.admin.addr} --port ${nodes.adminvm.config.givc.admin.port} --cacert ${nodes.hostvm.givc.host.tls.caCertPath} --cert ${nodes.hostvm.givc.host.tls.certPath} --key ${nodes.hostvm.givc.host.tls.keyPath} ${if tls then "" else "--notls"} --name ${nodes.adminvm.config.givc.admin.name} start --vm chromium-vm clearexit"))
                   time.sleep(20) # Give few seconds to application to spin up, exit, then start it again
                   print(hostvm.succeed("${cli} --addr ${nodes.adminvm.config.givc.admin.addr} --port ${nodes.adminvm.config.givc.admin.port} --cacert ${nodes.hostvm.givc.host.tls.caCertPath} --cert ${nodes.hostvm.givc.host.tls.certPath} --key ${nodes.hostvm.givc.host.tls.keyPath} ${if tls then "" else "--notls"} --name ${nodes.adminvm.config.givc.admin.name} start --vm chromium-vm clearexit"))
+              with subtest("VM restart"):
+                  appvm.crash()
+                  hostvm.succeed("systemctl stop [email protected]")
+                  time.sleep(20)
+                  appvm.start()
+                  time.sleep(20)
+                  print(hostvm.succeed("${cli} --addr ${nodes.adminvm.config.givc.admin.addr} --port ${nodes.adminvm.config.givc.admin.port} --cacert ${nodes.hostvm.givc.host.tls.caCertPath} --cert ${nodes.hostvm.givc.host.tls.certPath} --key ${nodes.hostvm.givc.host.tls.keyPath} ${if tls then "" else "--notls"} --name ${nodes.adminvm.config.givc.admin.name} start --vm chromium-vm foot"))
+                  assert hostvm.succeed("cat /tmp/[email protected]") == "2"
+                  wait_for_window("ghaf@appvm")
             '';
         };
       };