Closed GoogleCodeExporter closed 9 years ago
Watchdog is fine:
root 9405 1 0 12:27 ? 00:00:08 ruby
/opt/ovz-web-panel//utils/watchdog/watchdog.rb start
root 9412 1 0 12:27 ? 00:00:02 ruby
/opt/ovz-web-panel//script/server webrick -e production -p 3000 -b 0
root 9415 1 0 12:27 ? 00:00:13 ruby
/opt/ovz-web-panel//utils/hw-daemon/hw-daemon.rb start
the restart of watchdog didn't help
Original comment by s...@intellisoftware.net
on 24 May 2011 at 11:06
Can you provide error details saved in /opt/ovz-web-panel/log/production.log ?
Original comment by sibprogrammer
on 24 May 2011 at 4:06
Fixed after rebooting physical host.
I started Web Panel manually and issue gone.
Original comment by s...@intellisoftware.net
on 24 May 2011 at 5:25
Ok
Original comment by sibprogrammer
on 25 May 2011 at 3:03
[deleted comment]
Me too.
I have two almost similar boxes (debian.5, AMD x6, same mem, etc.). And one box
affected this problem - another is not.
I using owp since 1.7 and just ignore this problem until today. Now upgraded to
2.0 and problem still here.
One difference - affected box _is_ running ruby services in VE, another is not.
One more - it`s look like you can live longer after restart if you don`t try to
manipulate - just looking.
If change any VE then fall is near.
I played above an hour with 2.0 - all in attached log.
Original comment by flashv...@gmail.com
on 26 May 2011 at 9:35
Attachments:
same problem with OpenVZ Web Panel 2.0
Internal error: undefined method `pluralize' for #<Object:0xb680af3c>
Rendering /opt/ovz-web-panel/public/500.html (500 Internal Server Error)
restarting of OVZ-web-panel fixes but not for a long time
Original comment by shaki...@gmail.com
on 29 May 2011 at 11:45
Confirmed on Debian squeeze, using ovz-web-panel 2.0. It recently began
behaving as described, I can however not pinpoint the time to something
particular changed in the system or openvz setup.
Restarting watchdog DID help! Although only for a couple more requests.
Restarting watchdog again would continue resolve the issue temporarily.
Attaching log entries from the erroneous request.
Also, consider moving the project to another host. Perhaps one that does not
require you to use a Google Account just to report/comment on bus... ;-)
Original comment by k...@technopragmatics.org
on 5 Jun 2011 at 5:19
Attachments:
[deleted comment]
Got it!
In short - server fail, probably, when try to render zero value in server free
disk space bar.
To reproduce on 2.0
1. Take good working server, without any zero percents in df output
2. Mount iso image via loop device
3. Sync server via web
Result.
Server fail with Internal error: undefined method `pluralize' for
#<Object:0x7fb481bba5c0>
More digging.
First of all i look in the code of helper to look in pluralize invocation
vendor/rails/actionpack/lib/action_view/helpers/number_helper.rb
####
def number_to_human_size(number, *args)
return number.nil? ? nil : pluralize(number.to_i, "Byte") if number.to_i < 1024
####
I dig in ruby for 2 hours but still have no complete understand of this code.
(It is my first look in ruby :) )
So i look in code that call this helper
app/controllers/admin/hardware_servers_controller.rb
#####
:free => partition['free_bytes'],
:used => partition['used_bytes'],
:total => partition['total_bytes']
# :free => helper.number_to_human_size(partition['free_bytes'],
:locale => :en),
# :used => helper.number_to_human_size(partition['used_bytes'],
:locale => :en),
# :total =>
helper.number_to_human_size(partition['total_bytes'], :locale => :en)
###
I just tell that i agreed to get raw values instead of humanized and it is
solve the problem - not failing any more.
Later i connect host with patched code to another host that work stable and so
- used original code.
After sync - stable server fail - and at this point a start to think that
problem in data - not in code.
At this point i was really close to think about zero values. I mount loop
device and reproduce bug.
Hope it help someone - good luck!
Original comment by flashv...@gmail.com
on 7 Jun 2011 at 11:20
Original comment by sibprogrammer
on 18 Jun 2011 at 2:38
JFI: I got the same problem with the VPS pages itself
(/virtual-servers/show?id=1 etc) - log output attached.
Browser: Firefox 4.0.1 de_AT
OS:
Debian Squeeze, 64Bit ("Debian GNU/Linux 6.0 \n \l")
Versions:
Webpanel 2.0 (latest)
ruby 1.8.7 (2010-08-16 patchlevel 302) [x86_64-linux]
(installed over "apt-get install ruby rubygems libsqlite3-ruby libopenssl-ruby"
and all dependencys)
Original comment by wmaster20@gmail.com
on 18 Jun 2011 at 10:54
Attachments:
look for patch http://code.google.com/p/ovz-web-panel/issues/detail?id=248#c8
Original comment by flashv...@gmail.com
on 19 Jun 2011 at 3:53
Ah, thanks, that worked fine!
Original comment by wmaster20@gmail.com
on 19 Jun 2011 at 5:53
Guys, thank you very much for research and steps to reproduce.
Original comment by sibprogrammer
on 23 Jun 2011 at 4:55
This issue was closed by revision r913.
Original comment by sibprogrammer
on 23 Jun 2011 at 4:58
If possible try to apply patch from r913 and please confirm that everything is
ok.
Original comment by sibprogrammer
on 23 Jun 2011 at 4:59
Issue 381 has been merged into this issue.
Original comment by sibprogrammer
on 8 Dec 2011 at 3:46
Issue 248 has been merged into this issue.
Original comment by sibprogrammer
on 1 May 2012 at 2:56
Original issue reported on code.google.com by
s...@intellisoftware.net
on 24 May 2011 at 11:01