Handling user sessions?

Hi,

I notice that session data are stored locally under
RT_PATH/var/session_data. But there’s also an empty session table in the
RT database (we use Postgres).

In the future we need a system that can scale after need, and with
failover. So we need:

  • database and webserver on different machines.
  • a webserver cluster solution

Thus we need to find out how RT handles sessions data, so we can configure
a cluster solution.

  • database cluster/replication

The question may be; how can we change the default local file handling of
sessions data to use the sessions table instead, so we can put more than 1
webserver in front as a failover solution?

Tomas A. P. Olaj, email: tomas.olaj@usit.uio.no, web: folk.uio.no/tomaso
University of Oslo / USIT (Center for Information Technology Services)
System- and Application Management / Applications Management Group

Hi,

I notice that session data are stored locally under
RT_PATH/var/session_data. But there’s also an empty session table in the
RT database (we use Postgres).

Have you set RT::WebSessionClass? These days, RT should be defaulting
to sticking sessions in the database?

Jesse

On the marvelous Thu, 7 Apr 2005, Jesse Vincent wrote kindly to me …

I notice that session data are stored locally under
RT_PATH/var/session_data. But there’s also an empty session table in the
RT database (we use Postgres).

Have you set RT::WebSessionClass? These days, RT should be defaulting
to sticking sessions in the database?

Jesse

Thanks Jesse,

In the RT_SiteConfig.pm I see that the following is set:
Set($WebSessionClass , ‘Apache::Session::File’);

Should this be uncommented? The session table contains nothing, and
session info are stored locally.

Tomas A. P. Olaj, email: tomas.olaj@usit.uio.no, web: folk.uio.no/tomaso
University of Oslo / USIT (Center for Information Technology Services)
System- and Application Management / Applications Management Group

Hello once more,

More strange questions from me;

We notice that RT doesn’t include hard newlines e.g. ‘\n’ on messages sent
out from RT. Many e-mail clients doesn’t handle this “automagical” in
presentation, like Windows e-mail clients do.

I would guess that there is an internet standard (RFC) that recommends
max. chars per lines too?!

Tomas A. P. Olaj, email: tomas.olaj@usit.uio.no, web: folk.uio.no/tomaso
University of Oslo / USIT (Center for Information Technology Services)
System- and Application Management / Applications Management Group

Hello once more,

More strange questions from me;

We notice that RT doesn’t include hard newlines e.g. ‘\n’ on messages sent
out from RT. Many e-mail clients doesn’t handle this “automagical” in
presentation, like Windows e-mail clients do.

Again, this is somewhat a function of how you configure RT. Have a look
in the config file to see if you’ve changed these defaults:

Default width: 72

Set($MessageBoxWidth , 72);

Default wrapping: “HARD” (choices “SOFT”, “HARD”)

Set($MessageBoxWrap, “HARD”);

In the RT_SiteConfig.pm I see that the following is set:
Set($WebSessionClass , ‘Apache::Session::File’);

Should this be uncommented? The session table contains nothing, and
session info are stored locally.

That sure looks like you’ve locally told RT not to use Postgres to store
the sessions. You might want to talk to whomever did that.

On the marvelous Thu, 7 Apr 2005, Jesse Vincent wrote kindly to me …

Again, this is somewhat a function of how you configure RT. Have a look
in the config file to see if you’ve changed these defaults:

Default width: 72

Set($MessageBoxWidth , 72);

Default wrapping: “HARD” (choices “SOFT”, “HARD”)

Set($MessageBoxWrap, “HARD”);

You’ve right. I only responded on the behalf of our testers. My fault, not
checking that thorougly.

It comes clear to me that the “resolved”-template was written in a long
sentence. That may be the explaination.

Tomas A. P. Olaj, email: tomas.olaj@usit.uio.no, web: folk.uio.no/tomaso
University of Oslo / USIT (Center for Information Technology Services)
System- and Application Management / Applications Management Group

I would guess that there is an internet standard (RFC) that recommends
max. chars per lines too?!

RFC 2822 (the mail standard that replaced 822, and to which mailgate
ought to conform) recommends that lines hard-wrap at 78 characters, and
requires that they do so before (I think it’s) 998 characters. On the
network, that’s with a network EOL (CRLF), though it’s usually
represented differently internally depending on your platform.

Cheers,
– jra
Jay R. Ashworth jra@baylink.com
Designer Baylink RFC 2100
Ashworth & Associates The Things I Think '87 e24
St Petersburg FL USA http://baylink.pitas.com +1 727 647 1274

  If you can read this... thank a system administrator.  Or two.  --me

On the marvelous Thu, 7 Apr 2005, Jesse Vincent wrote kindly to me …

Again, this is somewhat a function of how you configure RT. Have a look
in the config file to see if you’ve changed these defaults:

Default width: 72

Set($MessageBoxWidth , 72);

Default wrapping: “HARD” (choices “SOFT”, “HARD”)

Set($MessageBoxWrap, “HARD”);

You’ve right. I only responded on the behalf of our testers. My fault, not
checking that thorougly.

It comes clear to me that the “resolved”-template was written in a long
sentence. That may be the explaination.

That raises a good, if possibly obvious point, Jesse. Those
parameters, I assume from their names, are not applied to templates,
right?

Cheers,
– jra
Jay R. Ashworth jra@baylink.com
Designer Baylink RFC 2100
Ashworth & Associates The Things I Think '87 e24
St Petersburg FL USA http://baylink.pitas.com +1 727 647 1274

  If you can read this... thank a system administrator.  Or two.  --me