However, if I stop mongrel manually and start it with the command "mongrel_rails -e production -p xxxx -d", it works as intended.
I'm wondering why the "restart" fails to do the job and the difference between the two.
Restarting Mongrel
- Thanna
- Posts: 3
- Starts: 3
- Wiki Edits: 0
- Location: Far East
Whenever Capistrano redeploys a new version of my app, I notice that the command "mongrel_rails restart" does not cause mongrel to serve up this newer version but instead keep serving the old version. Cap does report that a USR2 signal has been sent to Mongrel and the PID is changed as a result.
However, if I stop mongrel manually and start it with the command "mongrel_rails -e production -p xxxx -d", it works as intended.
I'm wondering why the "restart" fails to do the job and the difference between the two.
However, if I stop mongrel manually and start it with the command "mongrel_rails -e production -p xxxx -d", it works as intended.
I'm wondering why the "restart" fails to do the job and the difference between the two.
- William
- Posts: 1052
- Starts: 32
- Wiki Edits: 56
"restart" should theoretically issue a stop and then start. If you replace in a "mongrel_rails stop && mongrel_rails -e production -p xxxx -d" in your task does that work? What code are you using in your Capfile, exactly?
2007-08-26 10:11 AM
- Thanna
-
- Posts: 3
- Starts: 3
- Wiki Edits: 0
- Location: Far East
Hi William,
Not working...
task :restart, :roles => :app do
run "cd #{current_path} && mongrel_rails restart"
end
Working...
task :restart, :roles => :app do
run "cd #{current_path} && mongrel_rails stop"
run "cd #{current_path} && mongrel_rails start -e production -p xxxx -d"
end
Thank you.
Not working...
task :restart, :roles => :app do
run "cd #{current_path} && mongrel_rails restart"
end
Working...
task :restart, :roles => :app do
run "cd #{current_path} && mongrel_rails stop"
run "cd #{current_path} && mongrel_rails start -e production -p xxxx -d"
end
Thank you.
2007-08-26 11:21 PM
- William
-
- Posts: 1052
- Starts: 32
- Wiki Edits: 56
I remember hearing stories about this in the early days of Mongrel, but I had thought the newer versions had fixed the restart issues. Its possible there could be something in your app, or perhaps the amount/type of traffic you're getting, that's preventing it from working. We haven't had a report of this problem in over a year, however. I suppose the quick fix may be the easiest.
2007-08-27 12:40 AM
- William
-
- Posts: 1052
- Starts: 32
- Wiki Edits: 56
Actually- with recent version of Mongrel running on Ruby 1.8.6, that cgi_multipart EOF fix is not needed. Its possible that when cap issues the "cd #{current_path} && mongrel_rails restart" it gets confused by that warning message. You can test it by just issuing a mongrel_rails restart yourself and seeing what happens.
2007-08-27 04:43 AM
- Thanna
-
- Posts: 3
- Starts: 3
- Wiki Edits: 0
- Location: Far East
Hi,
Same prob issuing a manual mongrel_rails restart.
Anyway I'm going for the quick fix as this is not a big issue. But "restart" is more elegant if it works.
Thanks.
Same prob issuing a manual mongrel_rails restart.
Anyway I'm going for the quick fix as this is not a big issue. But "restart" is more elegant if it works.
Thanks.
2007-08-27 07:25 AM
- Ampp
- Posts: 1
- Starts: 0
- Wiki Edits: 0
Hello,
Just wanted to add that I experienced the exact same problem: "mongrel_rails restart" doesn't run the new code, but "mongrel_rails stop && mongrel_rails start -e production -p xxxx -d" does.
Just wanted to add that I experienced the exact same problem: "mongrel_rails restart" doesn't run the new code, but "mongrel_rails stop && mongrel_rails start -e production -p xxxx -d" does.