Open zhangwei1984 opened 8 years ago
Hello
About the chaining of VMs, when we wrote the ClickOS paper we used our own module for vale
that allowed us to "hardwire" ports to form the chain. However that module is not open source. One quick alternative is to use multiple bridges, each with the pair of ports you want to connect.
About cosmos
. Please use ClickOS-ctl. Cosmos is not compatible any more with the latest ClickOS version.
So, when mentioning latest ClickOS version, does it mean clickos v0.2 ?
clickos-ctl does not work. Eg, if I start an vale-based vm
root@nimbnode21:/home/wei# xl list
Name ID Mem VCPUs State Time(s)
Domain-0 0 32271 24 r----- 2072.1
clickos 18 8 1 -b---- 0.0
Then I use clickos-ctl launch vm, It shows invalid parameter
root@nimbnode21:~/clickos-ctl# ./clickos -m xenstore install -s 18 ../mirror.click
Invalid number of arguments for command install
root@nimbnode21:~/clickos-ctl# ./clickos -m xenstore install -s clickos ../mirror.click
Invalid number of arguments for command install
Hi,
Indeed, there is a fix already available as a pull request https://github.com/cnplab/clickos-ctl/pull/2 but the easiest solution would be remove the -m xenstore
given is anyway the default.
ClickOS control only works with the master branch of ClickOS.
I use the master branch generate the img. But I have the same exact problem which my vm crash once I launch the app. If I do not launch the click app,
root@nimbnode21:~# xl list
Name ID Mem VCPUs State Time(s)
Domain-0 0 32271 24 r----- 2100.1
clickos 19 8 1 r----- 13.9
root@nimbnode21:~/clickos-ctl# ./clickos install -s clickos ../mirror.click
[ clickos ][ 0 ]: installed new router.
[ clickos ][ 0 ]: started router.
[on_status:205] router id 0
[on_status:206] status change to Stopped
[on_status:205] router id 0
[on_status:206] status change to Running
Thread "click": pointer: 0x2000810a00, stack: 0x350000
backend dom 0
Mapping TX rings
0 map errors
Mapping RX rings
0 map errors
Mapping TX buffers
0 map errors
Mapping RX buffers
Cannot handle page request order 10!
0 map errors
root@nimbnode21:~/clickos-ctl# xl list
Name ID Mem VCPUs State Time(s)
Domain-0 0 32271 24 r----- 2101.0
clickos 19 8 1 ---sc- 15.4
Please try to assign more memory to the VM.
You mean in *.cfg file, assign , for e.g. memory = 512 (MB)?
On Thu, Mar 24, 2016 at 6:23 PM, Filipe Manco notifications@github.com wrote:
Please try to assign more memory to the VM.
— You are receiving this because you commented. Reply to this email directly or view it on GitHub https://github.com/cnplab/clickos/issues/24#issuecomment-201051538
V/R,
Wenhui Zhang
Email: wenhui@gwmail.gwu.edu Telephone: 1-(703) 424 3193
Yes, however you shouldn't need 512 MB. Usually it will be something like 16 MB. There is no method to determine how much memory is necessary, so you need to just try out.
No, it does not work. I changed it to 64, it still gets crashed. Other tries?
You shouldn't need that much memory, but please try with 512MB or 1GB.
Thanks @fmanco , awesome information~
Could anyone give some guidance how to do the VM service chain? The same experiment is clickos figure10. How should I write the click example, then make them chain together and the packets can go to the VM service chain?
Another problem is that when I compile clickos, it shows xenstore.h something header files which can not be found. After hardcoded it, it can get compiled successfully. The VM can start successfully.
But if it uses cosmos launch the click example, It does not show errors. But I can not see any happening on vm console side. It should print something. I am wondering whether other than hardcoded, should I do some extra configurations?