[pve-devel] [PATCH qemu-server v2] vm_start: add syslog info with which PID a VM was started

[pve-devel] [PATCH qemu-server v2] vm_start: add syslog info with which PID a VM was started

[Daniel Kral]

Adds a syslog entry to log the process id that has been given to the
QEMU VM process at start. This is helpful debugging information if the
pid shows up at other places, like a kernel stack trace, while the VM
has been running, but cannot be retrieved anymore (e.g. the pidfile has
been deleted or only the syslog is available).

The syslog has been put in the `PVE::QemuServer::vm_start_nolock`
subroutine to make sure that the PID is logged not only when the VM has
been started by the API endpoint `vm_start`, but also when the VM is
started by a remote migration.

Suggested-by: Hannes Dürr <h.duerr@proxmox.com>
Suggested-by: Friedrich Weber <f.weber@proxmox.com>
Signed-off-by: Daniel Kral <d.kral@proxmox.com>
---
Differences to v1:
    - Fix commit message trailer order
    - Uppercase PID in syslog message

 PVE/QemuServer.pm | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index b26da505..ec370473 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -43,6 +43,7 @@ use PVE::ProcFSTools;
 use PVE::PBSClient;
 use PVE::RESTEnvironment qw(log_warn);
 use PVE::RPCEnvironment;
+use PVE::SafeSyslog;
 use PVE::Storage;
 use PVE::SysFSTools;
 use PVE::Systemd;
@@ -5971,6 +5972,8 @@ sub vm_start_nolock {
     eval { PVE::QemuServer::PCI::reserve_pci_usage($pci_reserve_list, $vmid, undef, $pid) };
     warn $@ if $@;
 
+    syslog("info", "VM $vmid started with PID $pid.");
+
     if (defined(my $migrate = $res->{migrate})) {
 	if ($migrate->{proto} eq 'tcp') {
 	    my $nodename = nodename();
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

Re: [pve-devel] [PATCH qemu-server v2] vm_start: add syslog info with which PID a VM was started

[Daniel Herzig]

This works flawlessly on my testsetup. I find this information very
useful.

Reviewed-by: Daniel Herzig <d.herzig@proxmox.com>

Daniel Kral <d.kral@proxmox.com> writes:

> Adds a syslog entry to log the process id that has been given to the
> QEMU VM process at start. This is helpful debugging information if the
> pid shows up at other places, like a kernel stack trace, while the VM
> has been running, but cannot be retrieved anymore (e.g. the pidfile has
> been deleted or only the syslog is available).
>
> The syslog has been put in the `PVE::QemuServer::vm_start_nolock`
> subroutine to make sure that the PID is logged not only when the VM has
> been started by the API endpoint `vm_start`, but also when the VM is
> started by a remote migration.
>
> Suggested-by: Hannes Dürr <h.duerr@proxmox.com>
> Suggested-by: Friedrich Weber <f.weber@proxmox.com>
> Signed-off-by: Daniel Kral <d.kral@proxmox.com>
> ---
> Differences to v1:
>     - Fix commit message trailer order
>     - Uppercase PID in syslog message
>
>  PVE/QemuServer.pm | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
> index b26da505..ec370473 100644
> --- a/PVE/QemuServer.pm
> +++ b/PVE/QemuServer.pm
> @@ -43,6 +43,7 @@ use PVE::ProcFSTools;
>  use PVE::PBSClient;
>  use PVE::RESTEnvironment qw(log_warn);
>  use PVE::RPCEnvironment;
> +use PVE::SafeSyslog;
>  use PVE::Storage;
>  use PVE::SysFSTools;
>  use PVE::Systemd;
> @@ -5971,6 +5972,8 @@ sub vm_start_nolock {
>      eval { PVE::QemuServer::PCI::reserve_pci_usage($pci_reserve_list, $vmid, undef, $pid) };
>      warn $@ if $@;
>  
> +    syslog("info", "VM $vmid started with PID $pid.");
> +
>      if (defined(my $migrate = $res->{migrate})) {
>  	if ($migrate->{proto} eq 'tcp') {
>  	    my $nodename = nodename();


_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel