search

columbia / crane

Transparent Paxos
44 stars 16 forks source link

hardcoded Columbia IPs in new-run.sh #7

Closed vschiavoni closed 8 years ago

vschiavoni commented 8 years ago

Currently the new-run.sh file has hardcoded hostnames pointing to Columia machines:

# Update worker-run.py to the server
scp worker-run.py bug03.cs.columbia.edu:~/
scp worker-run.py bug01.cs.columbia.edu:~/
scp worker-run.py bug02.cs.columbia.edu:~/

How should we change those ? Without modifying those, executing the new-run one gets these:

Permission denied (publickey).
lost connection
Permission denied (publickey).
lost connection
Permission denied (publickey).
lost connection
Permission denied (publickey).
lost connection
vschiavoni commented 8 years ago

I've modified bug03 to point to u3 (my local LXC container running apache), bug01 to u1, and bug02 to u2. Is this the expected pattern ?

hemingcui commented 8 years ago

Dear Valerio, I recommend you keep using u1 on different replica, as our checkpoint scripts use u1 as the container name. Keeping the same container name is better for checkointing and restoring on different machines. I think a good solution is use your own IPs instead of bug0x.cs.columbia.edu as the IPs.

On Thu, May 12, 2016 at 11:01 PM, Valerio Schiavoni < notifications@github.com> wrote:

I've modified bug03 to point to u3 (my local LXC container running apache), bug01 to u1, and bug02 to u2. Is this the expected pattern ?

— You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub https://github.com/columbia/crane/issues/7#issuecomment-218785073

vschiavoni commented 8 years ago

Thanks for the suggestion. This can be closed.