Hao Yu
Commercial Scale Out
IBM T.J. Watson Research Center
1101 Kitchawan Rd/Route 134, Rm 36-019
Yorktown Heights, NY 10598-0218
914-945-1854, Email: yuh@us.ibm.com
----- Forwarded by Hao Yu/Watson/IBM on 10/31/2006 09:52 AM -----
Jimi
Xenidis/Watson/IB
M To
Hao Yu/Watson/IBM@IBMUS
10/31/2006 09:36 cc
AM Dan E Poff/Watson/IBM@IBMUS, David
M Daly/Watson/IBM@IBMUS, Dilma M Da
silva/Watson/IBM@IBMUS, Doron
Shiloach/Watson/IBM@IBMUS, Edi
Shmueli/Watson/IBM@IBMUS, Jack
Kouloheris/Watson/IBM@IBMUS,
Jiuxing Liu/Watson/IBM@IBMUS,
Jonghyuk Choi/Watson/IBM@IBMUS
Subject
Re: domU network problem , 10/30
progress(Document link: Hao Yu)
PLEASE, post these to the mailing list, it is important that the community
sees these issues.
I''m happy you have some networking, I have some more network stability
fixes that I''ll be pushing shortly.
-JX
Hao Yu/Watson/IBM
10/30/06 05:16 PM To
Jimi Xenidis/Watson/IBM@IBMUS
cc
David M Daly/Watson/IBM@IBMUS, Edi
Shmueli/Watson/IBM@IBMUS, Jonghyuk
Choi/Watson/IBM@IBMUS, Dan E
Poff/Watson/IBM@IBMUS, Jiuxing
Liu/Watson/IBM@IBMUS, Doron
Shiloach/Watson/IBM@IBMUS, Dilma M
Da silva/Watson/IBM@IBMUS, Jack
Kouloheris/Watson/IBM@IBMUS
Subject
domU network problem , 10/30
progress(Document link: Jimi
Xenidis)
Hi Jimi,
Dan, Jiuxing and I worked together on the problem. First we have following
summaries
1. the problem is now ''domU network not working properly'' :
ping-to-outside runs into error after 30 seconds ; domU hangs in nc/netcat
test when receiving long stream (MB or so). For the pint-to-outside issue,
when domU runs into error, outside can still ping to the domU.
cso238:/ # ping 9.2.78.83
PING 9.2.78.83 (9.2.78.83) 56(84) bytes of data.
64 bytes from 9.2.78.83: icmp_seq=1 ttl=64 time=28.8 ms
64 bytes from 9.2.78.83: icmp_seq=2 ttl=64 time=4.24 ms
64 bytes from 9.2.78.83: icmp_seq=3 ttl=64 time=3.81 ms
64 bytes from 9.2.78.83: icmp_seq=4 ttl=64 time=10.0 ms
64 bytes from 9.2.78.83: icmp_seq=5 ttl=64 time=5.05 ms
64 bytes from 9.2.78.83: icmp_seq=6 ttl=64 time=9.65 ms
64 bytes from 9.2.78.83: icmp_seq=7 ttl=64 time=10.0 ms
64 bytes from 9.2.78.83: icmp_seq=8 ttl=64 time=10.0 ms
64 bytes from 9.2.78.83: icmp_seq=9 ttl=64 time=3.96 ms
64 bytes from 9.2.78.83: icmp_seq=10 ttl=64 time=10.0 ms
64 bytes from 9.2.78.83: icmp_seq=11 ttl=64 time=3.89 ms
64 bytes from 9.2.78.83: icmp_seq=12 ttl=64 time=2.73 ms
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
2. brctl does not work properly on our environment. ''strace brctl
show''
indicates ''ioctl(3, SIOCGIFBR, 0xffc381d4) = -1 EINVAL (Invalid
argument)'' . And on SLES 10 environment, brctl show works fine.
In addition, how the dhcp problem is gone ? Bulent reconfiged the dhcp
server to have more show-live IP available . Now, domU can get IP
cso238:/ # ifconfig
eth0 Link encap:Ethernet HWaddr 00:16:3E:F9:BB:B3
inet addr:9.2.78.238 Bcast:9.2.78.255 Mask:255.255.255.0
UP BROADCAST NOTRAILERS RUNNING MULTICAST MTU:1500 Metric:1
RX packets:39 errors:0 dropped:0 overruns:0 frame:0
TX packets:19 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:3884 (3.7 Kb) TX bytes:2508 (2.4 Kb)
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:20 errors:0 dropped:0 overruns:0 frame:0
TX packets:20 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:1256 (1.2 Kb) TX bytes:1256 (1.2 Kb)
We thought the ''ping error'' problem seemed non trivial.
Regards,
Hao
Hao Yu
Commercial Scale Out
IBM T.J. Watson Research Center
1101 Kitchawan Rd/Route 134, Rm 36-019
Yorktown Heights, NY 10598-0218
914-945-1854, Email: yuh@us.ibm.com
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users
Hi,
This is just to report that the problem of ''brctl'' is fixed by
installing
the latest bridge-utils rpm : bridge-utils-1.2-5.ppc64.rpm .
With the latest bridge-utils installed, I
. used the original /etc/xen/scripts/xen-network-common.sh
. changed the vif line in the domU configuration file to "vif = [
'''' ]"
domU is booted properly with dhcp working. the ''brctl show'',
''brctl
showmacs xenbr1'' kind of commands give similar output as the reference
machine (Dans js20 running sles 10).
However, the ping-out still gives ''no buffer space available''
after exact
12 successfuly echo-replies.
Details:
. On dan''s machine, the bridge-utils version, based rpm -qa, is
bridge-utils-1.0.6-14.1 . I could not find the exact package through
rpmfind. I used the latest version . Here is the binary package : (See
attached file: bridge-utils-1.2-5.ppc64.rpm)
. David, today, on blade cso83, GPFS directory is mounting during booting.
Dilma looked at it. Now we need to (1) kill the runmmfs process first, and
then (2) run mmstartup to mount /gpfs1xen . This maybe due to your change
to the cso script as you mentioned ?
Regards,
Hao
Hao Yu
Commercial Scale Out
IBM T.J. Watson Research Center
1101 Kitchawan Rd/Route 134, Rm 36-019
Yorktown Heights, NY 10598-0218
914-945-1854, Email: yuh@us.ibm.com
Hao Yu/Watson/IBM
10/31/2006 09:53 To
AM xen-users@lists.xensource.com,
xen-devel@lists.xensource.com
cc
Subject
Fw: domU network problem , 10/30
progress
Hao Yu
Commercial Scale Out
IBM T.J. Watson Research Center
1101 Kitchawan Rd/Route 134, Rm 36-019
Yorktown Heights, NY 10598-0218
914-945-1854, Email: yuh@us.ibm.com
----- Forwarded by Hao Yu/Watson/IBM on 10/31/2006 09:52 AM -----
Jimi
Xenidis/Watson/IB
M To
Hao Yu/Watson/IBM@IBMUS
10/31/2006 09:36 cc
AM Dan E Poff/Watson/IBM@IBMUS, David
M Daly/Watson/IBM@IBMUS, Dilma M Da
silva/Watson/IBM@IBMUS, Doron
Shiloach/Watson/IBM@IBMUS, Edi
Shmueli/Watson/IBM@IBMUS, Jack
Kouloheris/Watson/IBM@IBMUS,
Jiuxing Liu/Watson/IBM@IBMUS,
Jonghyuk Choi/Watson/IBM@IBMUS
Subject
Re: domU network problem , 10/30
progress(Document link: Hao Yu)
PLEASE, post these to the mailing list, it is important that the community
sees these issues.
I''m happy you have some networking, I have some more network stability
fixes that I''ll be pushing shortly.
-JX
Hao Yu/Watson/IBM
10/30/06 05:16 PM To
Jimi Xenidis/Watson/IBM@IBMUS
cc
David M Daly/Watson/IBM@IBMUS, Edi
Shmueli/Watson/IBM@IBMUS, Jonghyuk
Choi/Watson/IBM@IBMUS, Dan E
Poff/Watson/IBM@IBMUS, Jiuxing
Liu/Watson/IBM@IBMUS, Doron
Shiloach/Watson/IBM@IBMUS, Dilma M
Da silva/Watson/IBM@IBMUS, Jack
Kouloheris/Watson/IBM@IBMUS
Subject
domU network problem , 10/30
progress(Document link: Jimi
Xenidis)
Hi Jimi,
Dan, Jiuxing and I worked together on the problem. First we have following
summaries
1. the problem is now ''domU network not working properly'' :
ping-to-outside runs into error after 30 seconds ; domU hangs in nc/netcat
test when receiving long stream (MB or so). For the pint-to-outside issue,
when domU runs into error, outside can still ping to the domU.
cso238:/ # ping 9.2.78.83
PING 9.2.78.83 (9.2.78.83) 56(84) bytes of data.
64 bytes from 9.2.78.83: icmp_seq=1 ttl=64 time=28.8 ms
64 bytes from 9.2.78.83: icmp_seq=2 ttl=64 time=4.24 ms
64 bytes from 9.2.78.83: icmp_seq=3 ttl=64 time=3.81 ms
64 bytes from 9.2.78.83: icmp_seq=4 ttl=64 time=10.0 ms
64 bytes from 9.2.78.83: icmp_seq=5 ttl=64 time=5.05 ms
64 bytes from 9.2.78.83: icmp_seq=6 ttl=64 time=9.65 ms
64 bytes from 9.2.78.83: icmp_seq=7 ttl=64 time=10.0 ms
64 bytes from 9.2.78.83: icmp_seq=8 ttl=64 time=10.0 ms
64 bytes from 9.2.78.83: icmp_seq=9 ttl=64 time=3.96 ms
64 bytes from 9.2.78.83: icmp_seq=10 ttl=64 time=10.0 ms
64 bytes from 9.2.78.83: icmp_seq=11 ttl=64 time=3.89 ms
64 bytes from 9.2.78.83: icmp_seq=12 ttl=64 time=2.73 ms
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
ping: sendmsg: No buffer space available
2. brctl does not work properly on our environment. ''strace brctl
show''
indicates ''ioctl(3, SIOCGIFBR, 0xffc381d4) = -1 EINVAL (Invalid
argument)'' . And on SLES 10 environment, brctl show works fine.
In addition, how the dhcp problem is gone ? Bulent reconfiged the dhcp
server to have more show-live IP available . Now, domU can get IP
cso238:/ # ifconfig
eth0 Link encap:Ethernet HWaddr 00:16:3E:F9:BB:B3
inet addr:9.2.78.238 Bcast:9.2.78.255 Mask:255.255.255.0
UP BROADCAST NOTRAILERS RUNNING MULTICAST MTU:1500 Metric:1
RX packets:39 errors:0 dropped:0 overruns:0 frame:0
TX packets:19 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:3884 (3.7 Kb) TX bytes:2508 (2.4 Kb)
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:20 errors:0 dropped:0 overruns:0 frame:0
TX packets:20 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:1256 (1.2 Kb) TX bytes:1256 (1.2 Kb)
We thought the ''ping error'' problem seemed non trivial.
Regards,
Hao
Hao Yu
Commercial Scale Out
IBM T.J. Watson Research Center
1101 Kitchawan Rd/Route 134, Rm 36-019
Yorktown Heights, NY 10598-0218
914-945-1854, Email: yuh@us.ibm.com
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users
Reasonably Related Threads
- Fw: domU network problem , 10/30 progress
- [LLVMdev] Issue with std::call_once in PPC64 platform
- [LLVMdev] FPOpFusion = Fast and Multiply-and-add combines
- [LLVMdev] Issue with std::call_once in PPC64 platform
- [LLVMdev] Issue with std::call_once in PPC64 platform