Hello emmanuel / all,

Thanks for the reply.. so you mean to say , connections had gone beyond the limit (greater than 1024(assuming))?.

And also (in the above case I explained) , when given a ctrl c signal, slapd did not shut down , it says waiting for (many) threads to terminate and hangs

Thanks,

Arunachalam

**************************************************************************** ****************************

This e-mail and attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained herein in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient's) is prohibited. If you receive this e-mail in error, please notify the sender by phone or email immediately and delete it!

-----Original Message----- From: openldap-software-bounces+arunachalamp=huawei.com@OpenLDAP.org [mailto:openldap-software-bounces+arunachalamp=huawei.com@OpenLDAP.org] On Behalf Of Emmanuel Dreyfus Sent: Monday, September 17, 2007 1:41 AM To: Arunachalam Parthasarathy; openldap-software@openldap.org Subject: Re: connection hangup -inference and solution?

Arunachalam Parthasarathy arunachalamp@huawei.com wrote:

Can I know , why it is happening, and whats the way to get rid of this

problem

I had similar problem because slapd did hit the system file descriptor

limit. It did not crash but stopped operating.

In my machine it is set to 1024, and to say , I had requested that many connections (used 50 clients maximum and delta syncrepl slave)

But still I am getting the error

**************************************************************************** ****************************

This e-mail and attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained herein in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient's) is prohibited. If you receive this e-mail in error, please notify the sender by phone or email immediately and delete it!

-----Original Message----- From: Emmanuel Dreyfus [mailto:manu@netbsd.org] Sent: Monday, September 17, 2007 3:35 PM To: Arunachalam Parthasarathy Cc: openldap-software@openldap.org Subject: Re: connection hangup -inference and solution?

On Mon, Sep 17, 2007 at 10:26:19AM +0530, Arunachalam Parthasarathy wrote:

Thanks for the reply.. so you mean to say , connections had gone beyond

the

limit (greater than 1024(assuming))?.

It's the system limit (ulimit -n), and the default was quite low in my

case: 64. Setting it to 512 fixes the problem, but your experience may

vary, as it depends on how many connexions you get.

Hello Quanah,

I didn't ran the ldap server under gdb , I ll do it with gdb for this same case and let you know,

But during this daemon hangup, it said hangup of descriptor value and after some time it said hangup on lot of descriptors

And to say , when it hanged during ctrl c , saying "waiting for threads to terminate", it was in futex wait in strace

Thanks,

Arunachalam.

**************************************************************************** ****************************

This e-mail and attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained herein in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient's) is prohibited. If you receive this e-mail in error, please notify the sender by phone or email immediately and delete it!

-----Original Message----- From: openldap-software-bounces+arunachalamp=huawei.com@OpenLDAP.org [mailto:openldap-software-bounces+arunachalamp=huawei.com@OpenLDAP.org] On Behalf Of Quanah Gibson-Mount Sent: Tuesday, September 18, 2007 3:15 AM To: openldap-software@openldap.org Subject: Re: connection hangup -inference and solution?

--On September 17, 2007 5:19:05 PM +0200 Tony Earnshaw tonni@hetnet.nl

wrote:

Arunachalam Parthasarathy skrev, on 16-09-2007 10:35:

...

I am using openldap 2.3.36 and bdb 4.5.20

Ok stuff. Updating OL will likely not help in this particular case.

I'd be curious to see a backtrace on the slapd process under gdb, if it was

built with debugging symbols.

--Quanah

--

Quanah Gibson-Mount

Principal Software Engineer

Zimbra, Inc

--------------------

Zimbra :: the leader in open source messaging and collaboration

Hello tony,

I am sorry , I did not get your question .. because , db did not get corrupted for the case I explained, server got hungup

Thanks,

Arunachalam.

**************************************************************************** ****************************

This e-mail and attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained herein in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient's) is prohibited. If you receive this e-mail in error, please notify the sender by phone or email immediately and delete it!

-----Original Message----- From: openldap-software-bounces+arunachalamp=huawei.com@OpenLDAP.org [mailto:openldap-software-bounces+arunachalamp=huawei.com@OpenLDAP.org] On Behalf Of Tony Earnshaw Sent: Monday, September 17, 2007 8:49 PM Cc: openldap-software@openldap.org Subject: Re: connection hangup -inference and solution?

Arunachalam Parthasarathy skrev, on 16-09-2007 10:35:

I am using openldap 2.3.36 and bdb 4.5.20

Ok stuff. Updating OL will likely not help in this particular case.

I am using delta syncrepl for synchronization and the configuration

files are same as the examples provided in the test folder of package

(for master and slave)

I started the master and added some entries (I had already some 50,000

entries)

I started the slave and it was synchronizing.

When I stopped the slave and client which was adding entries, in the

master *I am getting connection hangup on a descriptor and the master is

not responding to any client which are searching*

*Can I know , why it is happening, and whats the way to get rid of this

problem*

Been watching the thread for a couple or three of revs, but nowhere did

anyone match my experience with the same empirical experience.

In my case(s) it was caused by a corrupt (patched) 4.2.52 changelog DB.

Deleting the log after a stop slapd then restarting (so that the DB

automatically got rebuilt) fixed it. Runnings a (Buchan-adapted)

db_recover -c had had no effect at any stage.

Why the DB should have got corrupt in the first case, I can only

surmise. DB_CONFIG in that directory was reasonable, but consumers *did*

disappear for periods, due to system upgrade reboots before this occurred.

Systems were RHEL5 and FC6, OL versions were 2.3.37 and 2.3.38.

--Tonni