Re: [PATCH v3] libvirtd: Increase NL buffer size for lots of interface

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Jan 11, 2016 at 02:32:44PM -0500, Laine Stump wrote:
On 01/11/2016 05:44 AM, Martin Kletzander wrote:
On Mon, Jan 11, 2016 at 02:59:00PM +0800, Leno Hou wrote:
1. When switching CPUs to offline/online in a system more than 128 cpus
2. When using virsh to destroy domain in a system with more interface

All of above happens nl_recv returned with error: No buffer space
available.
This patch sets the socket buffer size to 128K and turns on
message peeking
for nl_recv,as this would solve this problem totally and permanetly.


So if none of the above is true/happening...

Signed-off-by: Leno Hou <houqy@xxxxxxxxxxxxxxxxxx>
Cc: Wenyi Gao <wenyi@xxxxxxxxxxxxxxxxxx>
CC: Laine Stump <laine@xxxxxxxxx>
CC: Michal Privoznik <mprivozn@xxxxxxxxxx>
---
src/util/virnetlink.c | 10 ++++++++++
1 file changed, 10 insertions(+)

diff --git a/src/util/virnetlink.c b/src/util/virnetlink.c
index 679b48e..ea65cbc 100644
--- a/src/util/virnetlink.c
+++ b/src/util/virnetlink.c
@@ -65,10 +65,12 @@ struct virNetlinkEventHandle {

# ifdef HAVE_LIBNL1
#  define virNetlinkAlloc nl_handle_alloc
+#  define virSocketSetBufferSize nl_set_buffer_size
#  define virNetlinkFree nl_handle_destroy
typedef struct nl_handle virNetlinkHandle;
# else
#  define virNetlinkAlloc nl_socket_alloc
+#  define virSocketSetBufferSize nl_socket_set_buffer_size
#  define virNetlinkFree nl_socket_free
typedef struct nl_sock virNetlinkHandle;
# endif
@@ -696,6 +698,14 @@ virNetlinkEventServiceStart(unsigned int
protocol, unsigned int groups)
       goto error_server;
   }

+    if (virSocketSetBufferSize(srv->netlinknh, 131702, 0) < 0) {
+        virReportSystemError(errno,
+                "%s",_("cannot set netlink socket buffer size to
128k"));
+        goto error_server;
+    }
+
+    nl_socket_enable_msg_peek(srv->netlinknh);
+

... shouldn't this be non-fatal just in case?

I at first agreed with this [*] if we just issue a warning and
continue we would have the least possibility of regression on older
systems (or maybe some odd/old system that didn't allow setting a 128k
buffer?). But on the other hand, I think the likelyhood of this is
very low, and if it *does* happen we (the developers/maintainers) want
to know about it. If there's a warning in a log file and libvirt
continues to operate, the user isn't likely to report it. If there is
an error message and something doesn't work, then we will definitely
hear about it. So I think this should remain as an error.


Well, we might want to know about it, I just don't see the benefit of
it.  Anyway, it's ok with me the way it is now as well.  I was just
wondering.

Have a nice day,
Martin

Attachment: signature.asc
Description: PGP signature

--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list

[Index of Archives]     [Virt Tools]     [Libvirt Users]     [Lib OS Info]     [Fedora Users]     [Fedora Desktop]     [Fedora SELinux]     [Big List of Linux Books]     [Yosemite News]     [KDE Users]     [Fedora Tools]