WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-api

[Xen-API] [PATCH 3 of 3] Startup function to execute an emergency networ

To: xen-api@xxxxxxxxxxxxxxxxxxx
Subject: [Xen-API] [PATCH 3 of 3] Startup function to execute an emergency network reset
From: Rob Hoes <rob.hoes@xxxxxxxxxx>
Date: Tue, 10 Aug 2010 17:10:31 +0100
Delivery-date: Tue, 10 Aug 2010 09:12:20 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <patchbomb.1281456628@xxxxxxxxxxxxxxxxxxxxx>
List-help: <mailto:xen-api-request@lists.xensource.com?subject=help>
List-id: Discussion of API issues surrounding Xen <xen-api.lists.xensource.com>
List-post: <mailto:xen-api@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-api>, <mailto:xen-api-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-api>, <mailto:xen-api-request@lists.xensource.com?subject=unsubscribe>
References: <patchbomb.1281456628@xxxxxxxxxxxxxxxxxxxxx>
Sender: xen-api-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: Mercurial-patchbomb/1.4.3
# HG changeset patch
# User Rob Hoes <rob.hoes@xxxxxxxxxx>
# Date 1281456025 -3600
# Node ID 576d2d14532b1a1764e6bc28bab21e48e7903ccc
# Parent  d15252d6d4137f7aaeffe61f537c336a5a4277e3
Startup function to execute an emergency network reset

This patch contains the code that is able to perform an emergency network 
reset. The code runs when xapi starts and the file /tmp/network-reset (created 
by the xe-reset-networking script) is present.

Signed-off-by: Rob Hoes <rob.hoes@xxxxxxxxxx>

diff -r d15252d6d413 -r 576d2d14532b ocaml/xapi/xapi.ml
--- a/ocaml/xapi/xapi.ml
+++ b/ocaml/xapi/xapi.ml
@@ -530,15 +530,56 @@
                        (Printexc.to_string e)
        end
 
-(* Make sure our license is set correctly *)
+(** Reset the networking-related metadata for this host if the command 
[xe-reset-networking]
+ *  was executed before the restart. *)
+let check_network_reset () =
+       try
+               (* Raises exception if the file is not there and no reset is 
required *)
+               let reset_file = Unixext.read_whole_file_to_string 
(Xapi_globs.network_reset_trigger) in
+               Server_helpers.exec_with_new_task "Performing emergency network 
reset"
+                       (fun __context ->
+                               let host = Helpers.get_localhost ~__context in
+                               (* Parse reset file *)
+                               let args = String.split '\n' reset_file in
+                               let args = List.map (fun s -> match 
(String.split '=' s) with k :: [v] -> k, v | _ -> "", "") args in
+                               let mAC = List.assoc "MAC" args in
+                               let mode = match List.assoc "MODE" args with
+                                       | "static" -> `Static
+                                       | "dhcp" | _ -> `DHCP
+                               in
+                               let iP = if List.mem_assoc "IP" args then 
List.assoc "IP" args else "" in
+                               let netmask = if List.mem_assoc "NETMASK" args 
then List.assoc "NETMASK" args else "" in
+                               let gateway = if List.mem_assoc "GATEWAY" args 
then List.assoc "GATEWAY" args else "" in
+                               let dNS = if List.mem_assoc "DNS" args then 
List.assoc "DNS" args else "" in
+                               
+                               (* Erase networking database objects for this 
host *)
+                               Helpers.call_api_functions ~__context
+                                       (fun rpc session_id ->
+                                               
Client.Client.Host.reset_networking rpc session_id host
+                                       );
+                               
+                               (* Introduce PIFs for remaining interfaces *)
+                               let pifs = Xapi_pif.scan_bios ~__context ~host 
in
+                               
+                               (* Introduce and configure the management PIF *)
+                               let pif = List.find (fun p -> Db.PIF.get_MAC 
~__context ~self:p = mAC) pifs in
+                               Xapi_pif.reconfigure_ip ~__context ~self:pif 
~mode ~iP ~netmask ~gateway ~dNS;
+                               Xapi_host.management_reconfigure ~__context 
~pif;
+                       );
+               (* Remove trigger file *)
+               Unix.unlink("/tmp/network-reset")
+       with _ -> () (* TODO: catch specific exception for missing fields in 
reset_file and inform user *)
+       
+
+(** Make sure our license is set correctly *)
 let handle_licensing () = 
-  Server_helpers.exec_with_new_task "Licensing host"
-    (fun __context ->
-       let host = Helpers.get_localhost ~__context in
-       License_init.initialise ~__context ~host;
-       (* Copy resulting license to the database *)
-       Xapi_host.copy_license_to_db ~__context ~host
-    )
+       Server_helpers.exec_with_new_task "Licensing host"
+               (fun __context ->
+                       let host = Helpers.get_localhost ~__context in
+                       License_init.initialise ~__context ~host;
+                       (* Copy resulting license to the database *)
+                       Xapi_host.copy_license_to_db ~__context ~host
+               )
 
 (** Writes the memory policy to xenstore and triggers the ballooning daemon. *)
 let control_domain_memory () =
@@ -765,62 +806,63 @@
   ];
     begin match Pool_role.get_role () with
     | Pool_role.Master ->
-       ()
+        ()
     | Pool_role.Broken ->
-       info "This node is broken; moving straight to emergency mode";
-       Xapi_host.set_emergency_mode_error Api_errors.host_broken [];
+        info "This node is broken; moving straight to emergency mode";
+        Xapi_host.set_emergency_mode_error Api_errors.host_broken [];
 
-       (* XXX: consider not restarting here *)
-       server_run_in_emergency_mode ()
+        (* XXX: consider not restarting here *)
+        server_run_in_emergency_mode ()
     | Pool_role.Slave _ ->
-       info "Running in 'Pool Slave' mode";
-       (* Set emergency mode until we actually talk to the master *)
-       Xapi_globs.slave_emergency_mode := true;
-       (* signal the init script that it should succeed even though we're bust 
*)
-       Helpers.touch_file !Xapi_globs.ready_file; 
-               
-       (* Keep trying to log into master *)
-       let finished = ref false in
-       while not(!finished) do
-         (* Grab the management IP address (wait forever for it if necessary) 
*)
-         let ip = wait_for_management_ip_address () in
-         
-         debug "Attempting to communicate with master";
-         (* Try to say hello to the pool *)
-         begin match attempt_pool_hello ip with
-         | None -> finished := true
-         | Some Temporary ->
-             debug "I think the error is a temporary one, retrying in 5s";
-             Thread.delay 5.;
-         | Some Permanent ->
-             error "Permanent error in Pool.hello, will retry after %.0fs just 
in case" Xapi_globs.permanent_master_failure_retry_timeout;
-             Thread.delay Xapi_globs.permanent_master_failure_retry_timeout
-         end;
-       done;
-       debug "Startup successful";
-       Xapi_globs.slave_emergency_mode := false;
-       Master_connection.connection_timeout := initial_connection_timeout;
-       
-       begin
-         try
-           (* We can't tolerate an exception in db synchronization so fall 
back into emergency mode
-              if this happens and try again later.. *)
-           Master_connection.restart_on_connection_timeout := false;
-           Master_connection.connection_timeout := 10.; (* give up retrying 
after 10s *)
-           Db_cache.DBCache.initialise_db_cache();
-           Dbsync.setup ()
-         with e ->
-           begin
-             debug "Failure in slave dbsync; slave will pause and then restart 
to try again. Entering emergency mode.";
-             server_run_in_emergency_mode()
-           end
-       end;
-       Master_connection.connection_timeout := 
Xapi_globs.master_connect_retry_timeout;
-       Master_connection.restart_on_connection_timeout := true;
-       Master_connection.on_database_connection_established := (fun () -> 
on_master_restart ~__context);
+        info "Running in 'Pool Slave' mode";
+        (* Set emergency mode until we actually talk to the master *)
+        Xapi_globs.slave_emergency_mode := true;
+        (* signal the init script that it should succeed even though we're 
bust *)
+        Helpers.touch_file !Xapi_globs.ready_file; 
+                
+        (* Keep trying to log into master *)
+        let finished = ref false in
+        while not(!finished) do
+          (* Grab the management IP address (wait forever for it if necessary) 
*)
+          let ip = wait_for_management_ip_address () in
+          
+          debug "Attempting to communicate with master";
+          (* Try to say hello to the pool *)
+          begin match attempt_pool_hello ip with
+          | None -> finished := true
+          | Some Temporary ->
+              debug "I think the error is a temporary one, retrying in 5s";
+              Thread.delay 5.;
+          | Some Permanent ->
+              error "Permanent error in Pool.hello, will retry after %.0fs 
just in case" Xapi_globs.permanent_master_failure_retry_timeout;
+              Thread.delay Xapi_globs.permanent_master_failure_retry_timeout
+          end;
+        done;
+        debug "Startup successful";
+        Xapi_globs.slave_emergency_mode := false;
+        Master_connection.connection_timeout := initial_connection_timeout;
+        
+        begin
+          try
+            (* We can't tolerate an exception in db synchronization so fall 
back into emergency mode
+               if this happens and try again later.. *)
+            Master_connection.restart_on_connection_timeout := false;
+            Master_connection.connection_timeout := 10.; (* give up retrying 
after 10s *)
+            Db_cache.DBCache.initialise_db_cache();
+            Dbsync.setup ()
+          with e ->
+            begin
+              debug "Failure in slave dbsync; slave will pause and then 
restart to try again. Entering emergency mode.";
+              server_run_in_emergency_mode()
+            end
+        end;
+        Master_connection.connection_timeout := 
Xapi_globs.master_connect_retry_timeout;
+        Master_connection.restart_on_connection_timeout := true;
+        Master_connection.on_database_connection_established := (fun () -> 
on_master_restart ~__context);
     end;
  
     Startup.run ~__context [
+      "Checking emergency network reset", [], check_network_reset;
       "Synchronising bonds/VLANs on slave with master", [], 
Sync_networking.sync_slave_with_master ~__context;
       "Initialise Monitor_rrds.use_min_max", [], 
Monitor_rrds.update_use_min_max;
       "Initialising licensing", [], handle_licensing;
diff -r d15252d6d413 -r 576d2d14532b ocaml/xapi/xapi_globs.ml
--- a/ocaml/xapi/xapi_globs.ml
+++ b/ocaml/xapi/xapi_globs.ml
@@ -610,5 +610,8 @@
 (** Default feature mask: EST (base_ecx.7) is ignored. *)
 let cpuid_default_feature_mask = "ffffff7f-ffffffff-ffffffff-ffffffff"
 
+(** Path to trigger file for Network Reset. *)
+let network_reset_trigger = "/tmp/network-reset"
+
 let first_boot_dir = "/etc/firstboot.d/"
 
 ocaml/xapi/xapi.ml       |  158 +++++++++++++++++++++++++++++-----------------
 ocaml/xapi/xapi_globs.ml |    3 +
 2 files changed, 103 insertions(+), 58 deletions(-)


Attachment: xen-api.hg-3.patch
Description: Text Data

_______________________________________________
xen-api mailing list
xen-api@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/mailman/listinfo/xen-api
<Prev in Thread] Current Thread [Next in Thread>