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

[Libguestfs] P2Vs seem to require a very robust Ethernet



Now that we can gather diagnostic info, I think I know why our P2Vs kept
failing last week.  Another one just died right in front of my eyes.  I
think either the Ethernet or NFS server at this site occasionally
"blips" offline when it gets busy and that messes up P2V migrations.
The RHEV export domain is an NFS share offered by an old Storagetek NAS,
connected over a 10/100 Ethernet.  All of us would prefer a gb Ethernet
but this is what we have available.  

First, an extract from /var/log/messages showing the start and end times
for virt-p2v-server, about 3 1/2 hours.

[root Fedora16-64P2V log]# tail /var/log/messages -c 1000
16-64P2V systemd-logind[704]: Removed session 16.
Nov 22 19:04:46 Fedora16-64P2V systemd-logind[704]: New session 17 of
user root.
Nov 22 19:04:48 Fedora16-64P2V virt-v2v[1733]: p2v-server started.
Nov 22 20:01:01 Fedora16-64P2V systemd-logind[704]: New session 18 of
user root.
Nov 22 20:01:02 Fedora16-64P2V systemd-logind[704]: Removed session 18.
Nov 22 21:01:01 Fedora16-64P2V systemd-logind[704]: New session 19 of
user root.
Nov 22 21:01:01 Fedora16-64P2V systemd-logind[704]: Removed session 19.
Nov 22 22:01:02 Fedora16-64P2V systemd-logind[704]: New session 20 of
user root.
Nov 22 22:01:02 Fedora16-64P2V systemd-logind[704]: Removed session 20.
Nov 22 22:33:44 Fedora16-64P2V virt-v2v[1733]: FATAL: Error receiving
data:
Nov 22 22:33:44 Fedora16-64P2V systemd-logind[704]: Removed session 17.
Nov 22 22:33:44 Fedora16-64P2V virt-v2v[1733]: WARNING: Error messages
were written to /var/log/virt-p2v-server.1322010288.log.
Nov 22 22:33:44 Fedora16-64P2V virt-v2v[1733]: p2v-server exited.
[root Fedora16-64P2V log]#


Next, the output from the trace log file from virt-p2v-server:

[root Fedora16-64P2V log]# more virt-p2v-server.1322010288.log
virt-v2v: Error receiving data:

[root Fedora16-64P2V log]#

Before it blew up, I was watching network usage of my RHEV host with the
Fedora migration server VM and it consistently pegged above 90%+.  I've
never seen an Ethernet that saturated before.  The iftop utility on my
Fedora guest migration server showed data consistently flying at 69-90+
Mb per second.  

- Greg Scott


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