[RAS] ReDIF-perl trouble
Sune Karlsson
Sune.Karlsson at oru.se
Mon Apr 11 05:40:13 UTC 2016
>From http://search.cpan.org/~miyagawa/Starman/script/starman
=====================================================================
HUP
Sending HUP signal to the master process will restart all the workers gracefully (meaning the currently running requests will shut down once the request is complete), and by default, the workers will pick up the code changes you make by reloading the application.
If you enable --preload-app option, however, the code will be only loaded in the startup process and will not pick up the code changes you made.
And
--preload-app
This option lets Starman preload the specified PSGI application in the master parent process before preforking children. This allows memory savings with copy-on-write memory management. When not set (default), forked children loads the application in the initialization hook.
Enabling this option can cause bad things happen when resources like sockets or database connections are opened at load time by the master process and shared by multiple children.
Since Starman 0.2000, this option defaults to false, and you should explicitly set this option to preload the application in the master process.
Alternatively, you can use -M command line option (plackup's common option) to preload the modules rather than the <application> itself.
starman -MCatalyst -MDBIx::Class myapp.psgi
will load the modules in the master process for memory savings with CoW, but the actual loading of myapp.psgi is done per children, allowing resource managements such as database connection safer.
If you enable this option, sending HUP signal to the master process will not pick up any code changes you make. See "SIGNALS" for details.
=========================================================
If I am reading this correctly you need to restart Starman without the --preload-app or -M options if they are used. You can then use HUP. To see the effect of code changes you should probably run it with as few workers as possible.
On the Email.pm module:
- I think the nasty hack comment refers to the line after the comment.
- I would try the built in debug/logging features. Enabling debug-email-data-log in the config, setting $debug to true in the code and instead of replacing the pipe to sendmail with opening a file adding something like
if ( $debug_email_data_log ) {
if ( open F, ">>:utf8", $debug_email_data_log ) {
print F scalar( localtime ), "\n";
print F $header, "\n", $body;
close F;
}
}
This way you can see both what happens to the e-mail and (hopefully) what sendmail gets.
Just my two cents.
/Sune
> -----Original Message-----
> From: 'Christian Zimmermann' [mailto:zimmermann at stlouisfed.org]
> Sent: Monday, April 11, 2016 2:19 AM
> To: Sune Karlsson
> Cc: RAS-run Mailing List
> Subject: Re: [RAS] ReDIF-perl trouble
>
> That is likely because starman is running the site, and I cannot find
> any
> documentation on how I can get it to source modified .pm files.
>
> On Sun, 10 Apr 2016, 'Christian Zimmermann' wrote:
>
> > And it turns out all my playing with the .pm is for naught, as they
> stay in
> > memory and are not read. What a nice Sunday I spent.
> >
> > Christian Zimmermann FIGUGEGL!
> > Economic Research
> > Federal Reserve Bank of St. Louis
> > P.O. Box 442
> > St. Louis MO 63166-0442 USA
> >
> https://mailfilter.sunet.se/canit/urlproxy.php?_q=aHR0cHM6Ly9pZGVhcy5yZX
> BlYy5vcmcvemltbS8%3D&_r=b3J1LXNl&_s=c2tuQG9ydS5zZQ%3D%3D
> @CZimm_economist
> >
> > On Sun, 10 Apr 2016, 'Christian Zimmermann' wrote:
> >
> >> There is
> >>
> >> print MESSAGE $header, "\n", $body;
> >>
> >>
> >> I have not yet been successful in printing to file. I have to wait
> for
> >> each email to reach max hops to see a result... Not the most elegant
> and
> >> quite frustrating.
> >>
> >> I am attaching what I think is the critical file, if that helps.
> >>
> >> Christian Zimmermann FIGUGEGL!
> >> Economic Research
> >> Federal Reserve Bank of St. Louis
> >> P.O. Box 442
> >> St. Louis MO 63166-0442 USA
> >>
> https://mailfilter.sunet.se/canit/urlproxy.php?_q=aHR0cHM6Ly9pZGVhcy5yZX
> BlYy5vcmcvemltbS8%3D&_r=b3J1LXNl&_s=c2tuQG9ydS5zZQ%3D%3D
> @CZimm_economist
> >>
> >> On Sun, 10 Apr 2016, Sune Karlsson wrote:
> >>
> >> > One thing that could be causing problems is if $header doesn't
> end with
> >> > a
> >> > newline. Then there would be no blank line between the headers
> and the
> >> > body of the message.
> >> >
> >> > I doubt that encode would do something really strange. Have you
> tried
> >> > printing to a file instead of piping?
> >> >
> >> > /Sune
> >> >
> >> > > -----Original Message-----
> >> > > From: 'Christian Zimmermann' [mailto:zimmermann at stlouisfed.org]
> >> > > Sent: Sunday, April 10, 2016 11:52 PM
> >> > > To: Dan Hayes
> >> > > Cc: Sune Karlsson; Thomas Krichel; RAS-run Mailing List
> >> > > Subject: Re: [RAS] ReDIF-perl trouble
> >> > >
> >> > > The configuration file for ACIS (the software for RAS) has:
> >> > >
> >> > > sendmail = '/usr/sbin/sendmail -t -f authors at repec.org'
> >> > >
> >> > > So it is not using a perl module to send email. It pipes to
> sendmail:
> >> > >
> >> > > my $sendmail = $config -> {sendmail};
> >> > > if ( open MESSAGE, "|-:utf8", $sendmail ) {
> >> > > print MESSAGE $header, "\n", $body;
> >> > > close MESSAGE;
> >> > > }
> >> > >
> >> > > I am currently playing with a line that has the following
> comment:
> >> > > my $val = encode( 'MIME-Q', $value );
> >> > > # a nasty hack to fix Encode's wrapping 'feature':
> >> > >
> >> > > Could be promising. It is applied to message headers.
> >> > >
> >> > > Christian Zimmermann FIGUGEGL!
> >> > > Economic Research
> >> > > Federal Reserve Bank of St. Louis
> >> > > P.O. Box 442
> >> > > St. Louis MO 63166-0442 USA
> >> > >
> https://mailfilter.sunet.se/canit/urlproxy.php?_q=aHR0cHM6Ly9pZGVhcy5yZX
> >> > > BlYy5vcmcvemltbS8%3D&_r=b3J1LXNl&_s=c2tuQG9ydS5zZQ%3D%3D
> >> > > @CZimm_economist
> >> > >
> >> > > On Sun, 10 Apr 2016, Dan Hayes wrote:
> >> > >
> >> > > > the mail gateway "mx1.stlouisfed.org" is configured in
> sendmail and
> >> > > postfix
> >> > > > config files, /etc/mail/sendmail.mc and /etc/postfix/main.cf.
> I'm
> >> > > > not
> >> > > sure
> >> > > > where it might be included in perl code or why it would be.
> >> > > >
> >> > > > On Sun, Apr 10, 2016 at 2:47 PM, Sune Karlsson
> >> > > > <Sune.Karlsson at oru.se>
> >> > > wrote:
> >> > > >
> >> > > > > Some thoughts off the top of my head. Most probably random
> noise
> >> > > > > or
> >> > > things
> >> > > > > you have already thought of.
> >> > > > >
> >> > > > > Is mx1.stlouisfed.org your mail gateway? Is this configured
> >> > > > > somewhere
> >> > > or
> >> > > > > picked up automatically? If configured it might help to
> trace
> >> > > > > where
> >> > > this
> >> > > > > variable is used.
> >> > > > >
> >> > > > > Perl module or system call for sending e-mail? I doubt that
> >> > > > > something
> >> > > is
> >> > > > > "messing" with the e-mail address. It is more likely that
> >> > > > > whatever is
> >> > > > > sending the e-mail tacks on the @mx1.stlouisfed.org to the
> >> > > > > address
> >> > > for
> >> > > > > some strange reason.
> >> > > > >
> >> > > > > /Sune
> >> > > > >
> >> > > > > > -----Original Message-----
> >> > > > > > From: 'Christian Zimmermann'
> [mailto:zimmermann at stlouisfed.org]
> >> > > > > > Sent: Sunday, April 10, 2016 8:03 PM
> >> > > > > > To: Sune Karlsson
> >> > > > > > Cc: Thomas Krichel; RAS-run Mailing List; Hayes, Dan --
> Dan
> >> > > > > > Hayes
> >> > > > > > Subject: RE: [RAS] ReDIF-perl trouble
> >> > > > > >
> >> > > > > > The latest of the little I was able to figure out.
> >> > > > > >
> >> > > > > > ReDIF-perl does not seem to be to blame, at least not
> directly.
> >> > > Updating
> >> > > > > > it triggered all sort of other updates, and despite
> reverting
> >> > > > > > ReDIF-
> >> > > > > > perl,
> >> > > > > > emails are *not* going out. New registrants and email
> changes
> >> > > bounced
> >> > > > > > hopping within our system.
> >> > > > > >
> >> > > > > > Emails are typically formatted like
> >> > > > > > zimmermann at stlouisfed.org@mx1.stlouisfed.org
> >> > > > > >
> >> > > > > > but I have seen one split in two like this:
> >> > > > > > christian.zimmermann at stls.frb.o@mx1.stlouisfed.org
> >> > > > > > rg at mx1.stlouisfed.org
> >> > > > > >
> >> > > > > > which makes absolutely no sense.
> >> > > > > >
> >> > > > > > There must be some package that has started messing with
> those
> >> > > > > > email
> >> > > > > > addresses. From the list of updated packages, there is no
> >> > > > > > obvious
> >> > > > > > candidate.
> >> > > > > >
> >> > > > > > Still digging.
> >> > > > > >
> >> > > > > > And, no, Thomas I cannot give you access to the machine.
> >> > > > > >
> >> > > > > >
> >> > > > > >
> >> > > > > > Christian Zimmermann FIGUGEGL!
> >> > > > > > Economic Research
> >> > > > > > Federal Reserve Bank of St. Louis
> >> > > > > > P.O. Box 442
> >> > > > > > St. Louis MO 63166-0442 USA
> >> > > > > >
> >> > >
> https://mailfilter.sunet.se/canit/urlproxy.php?_q=aHR0cHM6Ly9pZGVhcy5yZX
> >> > > > > > BlYy5vcmcvemltbS8%3D&_r=b3J1LXNl&_s=c2tuQG9ydS5zZQ%3D%3D
> >> > > > > > @CZimm_economist
> >> > > > > >
> >> > > > > > On Sun, 10 Apr 2016, Sune Karlsson wrote:
> >> > > > > >
> >> > > > > > > > > 2) I notice today the ReDIF-perl was not installed
> on
> >> > > > > > > > > RAS. after
> >> > > > > > > > > installation though, emails have broken. the to
> field is
> >> > > malformed,
> >> > > > > > > > for
> >> > > > > > > > > example:
> >> > > > > > > > >
> >> > > > > > > > > The following message to <"Christian Zimmermann"
> >> > > > > > > > > <zimmermann at stlouisfed.org>@mx1.stlouisfed.org> was
> >> > > > > > > > > undeliverable.
> >> > > > > > > > >
> >> > > > > > > > > returning to 2.70 fixed the problem, so it has
> nothing to
> >> > > > > > > > > do with
> >> > > > > > any
> >> > > > > > > > > external perl module that refreshed during install.
> I
> >> > > > > > > > > suspect
> >> > > this
> >> > > > > > has
> >> > > > > > > > to do
> >> > > > > > > > > with the following in Parser.pm:
> >> > > > > > > > >
> >> > > > > > > > > < 'remove_newline_from_values' => 0,
> >> > > > > > > > > > #SK 'remove_newline_from_values' => 0,
> >> > > > > > > > > > Obsoleted by
> >> > > new
> >> > > > > > > > > Unicode processing, newlines are always removed
> >> > > > > > > > >
> >> > > > > > > > > or at least that is all I can see.
> >> > > > > > >
> >> > > > > > > There are many differences between 2.70 and 2.74. It's
> not
> >> > > > > > > clear to
> >> > > me
> >> > > > > > that they should cause problems with sending e-mails. At
> least
> >> > > > > > I
> >> > > have
> >> > > > > > not experienced any problems.
> >> > > > > > >
> >> > > > > > > Which process is sending the e-mail, where does it get
> the
> >> > > > > > > e-mail
> >> > > > > > address? Perl module for sending e-mail or system call to
> an
> >> > > external
> >> > > > > > process? I suppose an assumption of a newline (or white
> space)
> >> > > > > > at
> >> > > the
> >> > > > > > end of the e-mail address that no longer holds could
> cause
> >> > > > > > problems
> >> > > when
> >> > > > > > building a command line.
> >> > > > > > >
> >> > > > > > > /Sune
> >> > > > > > >
> >> > > > >
> >> > > >
> >> >
> >>
> >
More information about the RAS-run
mailing list