Re: (ITS#4429) (back-ldap?) slapd deadlock

[richton@nbcs.rutgers.edu]

I'm not really sure how to find "connection failure" from the back-ldap
perspective (is there some particular Debug() you're looking for?) All
that shows up to stats syslog:

Mar 16 13:35:01 conn=38 op=2 ldap_back_retry: retrying
URI="ldap://ldap.nbcs.rutgers.edu"; DN=""
Mar 16 13:56:00 conn=79 op=2 ldap_back_retry: retrying
URI="ldap://ldap.nbcs.rutgers.edu"; DN=""
Mar 16 14:17:01 conn=116 op=2 ldap_back_retry: retrying
URI="ldap://ldap.nbcs.rutgers.edu"; DN=""

but this isn't a cause of death.


It turns out I set this installation to -d -1 before I left on vacation
(presently on; sorry for lack of response time). Is there something in
particular to look for?


On Fri, 10 Mar 2006, Pierangelo Masarati wrote:

> On Thu, 2006-03-09 at 19:13 +0000, richton@nbcs.rutgers.edu wrote:
> > Full_Name: Aaron Richton
> > Version: 2.3.20
> > OS: Solaris 9
> > URL:
> > Submission from: (NULL) (128.6.31.135)
> >
> >
> > My back-ldap installation is still producing fun. Once again, the user-visible
> > symptom is that binds stopped. This is with patches proposed in ITS#4409.
>
> What is the root cause of the connection failure?  This is essential to
> reproduce the issue and understand what is going on with the proxy's
> code when retrying a broken connection.  As I stated during ITS#4409
> investigation and resolution, I've tried many different failure cases,
> and the only hosed code I found was related to concurrent tries to
> implicitly bind a newly created connection.  I'm not even sure this
> would possibly occur for real, as I had to force it from inside the
> debugger.
>
> p.
>
>
>
>
> Ing. Pierangelo Masarati
> Responsabile Open Solution
> OpenLDAP Core Team
>
> SysNet s.n.c.
> Via Dossi, 8 - 27100 Pavia - ITALIA
> http://www.sys-net.it
> ------------------------------------------
> Office:   +39.02.23998309
> Mobile:   +39.333.4963172
> Email:    pierangelo.masarati@sys-net.it
> ------------------------------------------
>
>