I think this is probably worth doing another v0.8 release, maybe next
week.  There's also a few npm updates, which make shrinkwrap work
nicely with git://server:path.git urls, and go about 10x faster due to
fixing an unnecessary lockfile collision.


On Sat, Jun 8, 2013 at 3:18 PM, Ben Noordhuis <[email protected]> wrote:
> On Sat, Jun 8, 2013 at 2:21 PM, Steve Freegard <[email protected]> wrote:
>> Hi,
>>
>> I've just upgraded two machines running Haraka from node v0.8.16 to v0.8.24,
>> both machines are running identical versions or Haraka with the same
>> configuration along with the same kernel versions.
>>
>> One machine (that happens to run under xen) works just fine, the application
>> starts and forks one child per cpu whereas on the other machine all of the
>> child processes crash with the following error:
>>
>> Jun  8 06:53:42 mta25 haraka[1977]: [CRIT] [-] [core] TypeError: Cannot read
>> property 'getsockname' of undefined
>> Jun  8 06:53:42 mta25 haraka[1977]: [CRIT] [-] [core]     at net.js:943:23
>> Jun  8 06:53:42 mta25 haraka[1977]: [CRIT] [-] [core]     at
>> Object.cluster._getServer [as 145:1] (cluster.js:555:5)
>> Jun  8 06:53:42 mta25 haraka[1977]: [CRIT] [-] [core]     at handleResponse
>> (cluster.js:149:41)
>> Jun  8 06:53:42 mta25 haraka[1977]: [CRIT] [-] [core]     at respond
>> (cluster.js:170:5)
>> Jun  8 06:53:42 mta25 haraka[1977]: [CRIT] [-] [core]     at handleMessage
>> (cluster.js:180:5)
>> Jun  8 06:53:42 mta25 haraka[1977]: [CRIT] [-] [core]     at
>> process.EventEmitter.emit (events.js:126:20)
>> Jun  8 06:53:42 mta25 haraka[1977]: [CRIT] [-] [core]     at handleMessage
>> (child_process.js:270:12)
>> Jun  8 06:53:42 mta25 haraka[1977]: [CRIT] [-] [core]     at
>> Pipe.channel.onread (child_process.js:295:9)
>> Jun  8 06:53:42 mta25 haraka[1977]: [NOTICE] [-] [core] Shutting down
>>
>> I've tried reducing the number of child processes, but nothing works - I
>> can't get it to start under cluster at all.
>>
>> I've found https://github.com/joyent/node/issues/5330 - but that seems to
>> indicate that it affects node >=v0.10 only?
>> I also don't understand why it only affects one machine and not the other
>> (the affected machines does have more CPUs though).
>>
>> What can I do to avoid this bug or workaround it in the application?
>>
>> Kind regards,
>> Steve.
>
> I've back-ported the fix to the v0.8 branch in [1].  v0.8 releases are
> infrequent (it's in maintenance mode) so you may want to consider
> building from v0.8 HEAD.
>
> [1] https://github.com/joyent/node/commit/8a3d0c8
>
> --
> --
> Job Board: http://jobs.nodejs.org/
> Posting guidelines: 
> https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines
> You received this message because you are subscribed to the Google
> Groups "nodejs" group.
> To post to this group, send email to [email protected]
> To unsubscribe from this group, send email to
> [email protected]
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>
> ---
> You received this message because you are subscribed to the Google Groups 
> "nodejs" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to [email protected].
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

-- 
-- 
Job Board: http://jobs.nodejs.org/
Posting guidelines: 
https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines
You received this message because you are subscribed to the Google
Groups "nodejs" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"nodejs" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to