Fedora directory users - Oct 2006 - max file descriptor setting in fds

In the console I noticed a setting for max number of
file descriptors.
configuration > performance tab
Max number of file descriptors: 1024

Should I leave it at the default of 1024? Or should I
try to match what I did in performance tuning that was
on the page below? Maybe theyre not even related. I
just dont want to have conflicting settings.
http://directory.fedora.redhat.com/wiki/Performance_Tuning


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com

I have four machines in an MMR setup.  One recently died (got
accidentally unplugged; d''oh!).  When I started it back up, it
recovered the database and then dumped errors like the following in
the errors log:

[04/Oct/2006:13:59:03 -0500] NSMMReplicationPlugin -
agmt="cn="Replication to zeppo.nebrwesleyan.edu""
(zeppo:389): Missing
data encountered
[04/Oct/2006:13:59:04 -0500] NSMMReplicationPlugin -
agmt="cn="Replication to zeppo.nebrwesleyan.edu""
(zeppo:389):
Incremental update failed and requires administrator action

The only info I found on this via Google was for Sun DS, and it just
recommended updating to the latest hotfix.  Obviously not terribly
helpful.

Anyhow have any ideas how I might fix this, short of re-initializing
my directory on this machine?  Thanks!

I''m running:

Name        : fedora-ds                    Relocations: /opt/fedora-ds
Version     : 1.0.2                             Vendor: (none)
Release     : 1.RHEL4                       Build Date: Wed 01 Mar 2006 01:21:46
PM CST

I used the mmr.pl script to set up replication, and it''s been running
fine for about 9 months now.

Chris St. Pierre
Unix Systems Administrator
Nebraska Wesleyan University

Chris St. Pierre wrote:
> I have four machines in an MMR setup.  One recently died (got
> accidentally unplugged; d''oh!).  When I started it back up, it
> recovered the database and then dumped errors like the following in
> the errors log:
>
> [04/Oct/2006:13:59:03 -0500] NSMMReplicationPlugin -
> agmt="cn="Replication to zeppo.nebrwesleyan.edu""
(zeppo:389): Missing
> data encountered
> [04/Oct/2006:13:59:04 -0500] NSMMReplicationPlugin -
> agmt="cn="Replication to zeppo.nebrwesleyan.edu""
(zeppo:389):
> Incremental update failed and requires administrator action
>
> The only info I found on this via Google was for Sun DS, and it just
> recommended updating to the latest hotfix.  Obviously not terribly
> helpful.
>   
There is a very small window of time between when the update is written 
to the main database and when it is written to the changelog database.  
If you kill the power during this window, the changelog db will be out 
of sync with the main database.  The only solution is to reinit the 
server which has this problem.
> Anyhow have any ideas how I might fix this, short of re-initializing
> my directory on this machine?  Thanks!
>
> I''m running:
>
> Name        : fedora-ds                    Relocations: /opt/fedora-ds
> Version     : 1.0.2                             Vendor: (none)
> Release     : 1.RHEL4                       Build Date: Wed 01 Mar 2006
01:21:46 PM CST
>
> I used the mmr.pl script to set up replication, and it''s been
running
> fine for about 9 months now.
>
> Chris St. Pierre
> Unix Systems Administrator
> Nebraska Wesleyan University
>
> --
> Fedora-directory-users mailing list
> Fedora-directory-users@redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-directory-users
>

On Wed, 4 Oct 2006, Richard Megginson wrote:
> There is a very small window of time between when the update is written to
the
> main database and when it is written to the changelog database.  If you
kill
> the power during this window, the changelog db will be out of sync with the
> main database.  The only solution is to reinit the server which has this
> problem.
Thanks.  Next question: what''s the best (quickest, easiest) way to
reinitialize the server?  I can think of a few ways (e.g., reinstall),
but can''t shake the feeling that there must be a better way.

Chris St. Pierre
Unix Systems Administrator
Nebraska Wesleyan University

Chris St. Pierre wrote:
> On Wed, 4 Oct 2006, Richard Megginson wrote:
>
>   
>> There is a very small window of time between when the update is written
to the
>> main database and when it is written to the changelog database.  If you
kill
>> the power during this window, the changelog db will be out of sync with
the
>> main database.  The only solution is to reinit the server which has
this
>> problem.
>>     
>
> Thanks.  Next question: what''s the best (quickest, easiest) way to
> reinitialize the server?  I can think of a few ways (e.g., reinstall),
> but can''t shake the feeling that there must be a better way.
>   
Do you have another master?  If so, just perform a replica 
initialization (using the console).
> Chris St. Pierre
> Unix Systems Administrator
> Nebraska Wesleyan University
>
> --
> Fedora-directory-users mailing list
> Fedora-directory-users@redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-directory-users
>

On Thu, 5 Oct 2006, Richard Megginson wrote:
> Do you have another master?  If so, just perform a replica initialization
> (using the console).
We''ve got three other masters, but we don''t use the console --
or have
X installed or heads on the machines.  Can this be done from the CLI,
maybe with mmr.pl?

Chris St. Pierre
Unix Systems Administrator
Nebraska Wesleyan University

Chris St. Pierre wrote:
> On Thu, 5 Oct 2006, Richard Megginson wrote:
>
>   
>> Do you have another master?  If so, just perform a replica
initialization
>> (using the console).
>>     
>
> We''ve got three other masters, but we don''t use the
console -- or have
> X installed or heads on the machines.  Can this be done from the CLI,
> maybe with mmr.pl?
>   
Not sure about mmr.pl, but yes, this can be done via ldapmodify.
First, you have to find the DN of the replication agreement from another 
master to the one you want to initialize.
ldapsearch -x -h othermaster -D "cn=directory manager" -w password -s 
sub -b cn=config "objectclass=nsDS5ReplicationAgreement" cn 
nsDS5ReplicaHost nsDS5ReplicaPort
There may be several, you''ll have to figure out which one goes to the 
master you want to initialize

Next, use ldapmodify to start the repl init:
ldapmodify -x -h othermaster -D "cn=directory manager" -w password
dn: dn of the repl agreement
changetype: modify
replace: nsds5BeginReplicaRefresh
nsds5BeginReplicaRefresh: start

Then check the error logs for both masters to see when repl init is 
completed.
> Chris St. Pierre
> Unix Systems Administrator
> Nebraska Wesleyan University
>
> --
> Fedora-directory-users mailing list
> Fedora-directory-users@redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-directory-users
>

I went through and re-init both of the databases on the problem
machine, but still get the same errors.  Also, if I restart FDS on the
box, it claims to have been shutdown in a disorderly manner, even when
I use stop-slapd.  There are no errors related to the shutdown.

It sounds like the BDB backend is pretty hosed, and even the auto
recovery isn''t completely handling it.  Ideas?

Chris St. Pierre
Unix Systems Administrator
Nebraska Wesleyan University

On Thu, 5 Oct 2006, Richard Megginson wrote:
> Chris St. Pierre wrote:
>> On Thu, 5 Oct 2006, Richard Megginson wrote:
>>
>>   
>> > Do you have another master?  If so, just perform a replica
initialization
>> > (using the console).
>> >     
>>
>> We''ve got three other masters, but we don''t use the
console -- or have
>> X installed or heads on the machines.  Can this be done from the CLI,
>> maybe with mmr.pl?
>>   
> Not sure about mmr.pl, but yes, this can be done via ldapmodify.
> First, you have to find the DN of the replication agreement from another
master
> to the one you want to initialize.
> ldapsearch -x -h othermaster -D "cn=directory manager" -w
password -s sub -b
> cn=config "objectclass=nsDS5ReplicationAgreement" cn
nsDS5ReplicaHost
> nsDS5ReplicaPort
> There may be several, you''ll have to figure out which one goes to
the master
> you want to initialize
>
> Next, use ldapmodify to start the repl init:
> ldapmodify -x -h othermaster -D "cn=directory manager" -w
password
> dn: dn of the repl agreement
> changetype: modify
> replace: nsds5BeginReplicaRefresh
> nsds5BeginReplicaRefresh: start
>
> Then check the error logs for both masters to see when repl init is
completed.
>> Chris St. Pierre
>> Unix Systems Administrator
>> Nebraska Wesleyan University
>>
>> --
>> Fedora-directory-users mailing list
>> Fedora-directory-users@redhat.com
>> https://www.redhat.com/mailman/listinfo/fedora-directory-users
>>   
>

Chris St. Pierre wrote:
> I went through and re-init both of the databases on the problem
> machine, but still get the same errors.  Also, if I restart FDS on the
> box, it claims to have been shutdown in a disorderly manner, even when
> I use stop-slapd.  There are no errors related to the shutdown.
>
> It sounds like the BDB backend is pretty hosed, and even the auto
> recovery isn''t completely handling it.  Ideas?
>   
Does this happen every time with restarting the server, the disorderly 
shutdown in the errors log?  If so, does it recover the main database?  
If so, I think perhaps the changelog db is corrupted.  Try shutting down 
the server and removing the changelog db.
> Chris St. Pierre
> Unix Systems Administrator
> Nebraska Wesleyan University
>
> On Thu, 5 Oct 2006, Richard Megginson wrote:
>
>   
>> Chris St. Pierre wrote:
>>     
>>> On Thu, 5 Oct 2006, Richard Megginson wrote:
>>>
>>>   
>>>       
>>>> Do you have another master?  If so, just perform a replica
initialization
>>>> (using the console).
>>>>     
>>>>         
>>> We''ve got three other masters, but we don''t use
the console -- or have
>>> X installed or heads on the machines.  Can this be done from the
CLI,
>>> maybe with mmr.pl?
>>>   
>>>       
>> Not sure about mmr.pl, but yes, this can be done via ldapmodify.
>> First, you have to find the DN of the replication agreement from
another master
>> to the one you want to initialize.
>> ldapsearch -x -h othermaster -D "cn=directory manager" -w
password -s sub -b
>> cn=config "objectclass=nsDS5ReplicationAgreement" cn
nsDS5ReplicaHost
>> nsDS5ReplicaPort
>> There may be several, you''ll have to figure out which one goes
to the master
>> you want to initialize
>>
>> Next, use ldapmodify to start the repl init:
>> ldapmodify -x -h othermaster -D "cn=directory manager" -w
password
>> dn: dn of the repl agreement
>> changetype: modify
>> replace: nsds5BeginReplicaRefresh
>> nsds5BeginReplicaRefresh: start
>>
>> Then check the error logs for both masters to see when repl init is
completed.
>>     
>>> Chris St. Pierre
>>> Unix Systems Administrator
>>> Nebraska Wesleyan University
>>>
>>> --
>>> Fedora-directory-users mailing list
>>> Fedora-directory-users@redhat.com
>>> https://www.redhat.com/mailman/listinfo/fedora-directory-users
>>>   
>>>       
>
> --
> Fedora-directory-users mailing list
> Fedora-directory-users@redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-directory-users
>

If any of you are familiar with mailgraph for Postfix-based mail
servers, I''ve created something similar for Fedora DS.  fdsgraph tails
the access log and creates rrdtool-based graphs of the number of
connections and operations, organized by connection security and op
type respectively.  You can see some screenshots and find the tarball
at:

http://www.nebrwesleyan.edu/people/stpierre/fdsgraph/fdsgraph.html

Enjoy!

Chris St. Pierre
Unix Systems Administrator
Nebraska Wesleyan University

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Chris St. Pierre wrote:
> If any of you are familiar with mailgraph for Postfix-based mail
> servers, I''ve created something similar for Fedora DS.  fdsgraph
tails
> the access log and creates rrdtool-based graphs of the number of
> connections and operations, organized by connection security and op
> type respectively.  You can see some screenshots and find the tarball
> at:
> 
> http://www.nebrwesleyan.edu/people/stpierre/fdsgraph/fdsgraph.html
i''ve tried your tools but i''m having some error when starting
fdsgraph
service.
=====================[root@jstsvr fdsgraph]# service fdsgraph start
Starting fdsgraph
Can''t locate File/Tail.pm in @INC (@INC contains:
/usr/lib/perl5/site_perl/5.8.8/i386-linux-thread-multi
/usr/lib/perl5/site_perl/5.8.7/i386-linux-thread-multi
/usr/lib/perl5/site_perl/5.8.6/i386-linux-thread-multi
/usr/lib/perl5/site_perl/5.8.5/i386-linux-thread-multi
/usr/lib/perl5/site_perl/5.8.4/i386-linux-thread-multi
/usr/lib/perl5/site_perl/5.8.3/i386-linux-thread-multi
/usr/lib/perl5/site_perl/5.8.8 /usr/lib/perl5/site_perl/5.8.7
/usr/lib/perl5/site_perl/5.8.6 /usr/lib/perl5/site_perl/5.8.5
/usr/lib/perl5/site_perl/5.8.4 /usr/lib/perl5/site_perl/5.8.3
/usr/lib/perl5/site_perl
/usr/lib/perl5/vendor_perl/5.8.8/i386-linux-thread-multi
/usr/lib/perl5/vendor_perl/5.8.7/i386-linux-thread-multi
/usr/lib/perl5/vendor_perl/5.8.6/i386-linux-thread-multi
/usr/lib/perl5/vendor_perl/5.8.5/i386-linux-thread-multi
/usr/lib/perl5/vendor_perl/5.8.4/i386-linux-thread-multi
/usr/lib/perl5/vendor_perl/5.8.3/i386-linux-thread-multi
/usr/lib/perl5/vendor_perl/5.8.8 /usr/lib/perl5/vendor_perl/5.8.7
/usr/lib/perl5/vendor_perl/5.8.6 /usr/lib/perl5/vendor_perl/5.8.5
/usr/lib/perl5/vendor_perl/5.8.4 /usr/lib/perl5/vendor_perl/5.8.3
/usr/lib/perl5/vendor_perl /usr/lib/perl5/5.8.8/i386-linux-thread-multi
/usr/lib/perl5/5.8.8 .) at /usr/local/bin/fdsgraph.pl line 189.
BEGIN failed--compilation aborted at /usr/local/bin/fdsgraph.pl line 189.
========================
is there any package that i should install?

sigid
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFFNtHNqiPNNgPlDu0RAv+0AJ45cz6s81cC2qB8zhEaW5m6mHHvRACgmgp5
rUzqjejDsqDcVZI7F3WJI7Q=pSYk
-----END PGP SIGNATURE-----

sigid@JINLab escribió:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Chris St. Pierre wrote:
>   
>> If any of you are familiar with mailgraph for Postfix-based mail
>> servers, I''ve created something similar for Fedora DS. 
fdsgraph tails
>> the access log and creates rrdtool-based graphs of the number of
>> connections and operations, organized by connection security and op
>> type respectively.  You can see some screenshots and find the tarball
>> at:
>>
>> http://www.nebrwesleyan.edu/people/stpierre/fdsgraph/fdsgraph.html
>>     
> i''ve tried your tools but i''m having some error when
starting fdsgraph
> service.
> =====================> [root@jstsvr fdsgraph]# service fdsgraph start
> Starting fdsgraph
> Can''t locate File/Tail.pm in @INC (@INC contains:
> /usr/lib/perl5/site_perl/5.8.8/i386-linux-thread-multi
> /usr/lib/perl5/vendor_perl /usr/lib/perl5/5.8.8/i386-linux-thread-multi
> /usr/lib/perl5/5.8.8 .) at /usr/local/bin/fdsgraph.pl line 189.
> BEGIN failed--compilation aborted at /usr/local/bin/fdsgraph.pl line 189.
> ========================>
> is there any package that i should install?
>   
what about perl-File-Tail?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Patricio Bruna V. wrote:
> sigid@JINLab escribió:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> Chris St. Pierre wrote:
>>  
>>> If any of you are familiar with mailgraph for Postfix-based mail
>>> servers, I''ve created something similar for Fedora DS. 
fdsgraph tails
>>> the access log and creates rrdtool-based graphs of the number of
>>> connections and operations, organized by connection security and op
>>> type respectively.  You can see some screenshots and find the
tarball
>>> at:
>>>
>>> http://www.nebrwesleyan.edu/people/stpierre/fdsgraph/fdsgraph.html
>>>     
>> i''ve tried your tools but i''m having some error when
starting fdsgraph
>> service.
>> =====================>> [root@jstsvr fdsgraph]# service fdsgraph
start
>> Starting fdsgraph
>> Can''t locate File/Tail.pm in @INC (@INC contains:
>> /usr/lib/perl5/site_perl/5.8.8/i386-linux-thread-multi
>> /usr/lib/perl5/vendor_perl /usr/lib/perl5/5.8.8/i386-linux-thread-multi
>> /usr/lib/perl5/5.8.8 .) at /usr/local/bin/fdsgraph.pl line 189.
>> BEGIN failed--compilation aborted at /usr/local/bin/fdsgraph.pl line
189.
>> ========================>>
>> is there any package that i should install?
>>   
> what about perl-File-Tail?
hmm perl-File-tail is not installed yet.
but after i installed above packet, on browser doesn''t appear any graph
(only the header daily, weeky, so on)
btw before i start the browser when starting the fdsgraph after
installing perl-File-Tail, an error comes up

[root@jstsvr fdsgraph]# service fdsgraph start
Starting fdsgraph
fdsgraph: can''t chdir to /var/lib/fdsgraph: No such file or directory
at
/usr/local/bin/fdsgraph.pl line 251.

as an action i created the directory /var/lib/fdsgraph. then the service
could runs but there is no graph appears
am i missing some configuration? because i already follow the guide.

thanks
sigid
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFFOGUlqiPNNgPlDu0RAutBAKCtHblDDJiWv7AnnjNQcAq/z60Z2wCbBSRa
PBg0X0Rpgs7+NWWzOa8GdWw=vSoy
-----END PGP SIGNATURE-----

On Fri, 20 Oct 2006, sigid@JINLab wrote:
>hmm perl-File-tail is not installed yet.
>but after i installed above packet, on browser doesn''t appear any
graph
>(only the header daily, weeky, so on)
>btw before i start the browser when starting the fdsgraph after
>installing perl-File-Tail, an error comes up
>
>[root@jstsvr fdsgraph]# service fdsgraph start
>Starting fdsgraph
>fdsgraph: can''t chdir to /var/lib/fdsgraph: No such file or
directory at
>/usr/local/bin/fdsgraph.pl line 251.
>
>as an action i created the directory /var/lib/fdsgraph. then the service
>could runs but there is no graph appears
>am i missing some configuration? because i already follow the guide.
>
>thanks
>sigid
Check the ownership on /var/lib/fdsgraph.  It has to be readable by
your web server.

Also, check Apache''s error logs; you should find error messages in
there pointing you to what needs to be done.

Chris St. Pierre
Unix Systems Administrator
Nebraska Wesleyan University

Chris St. Pierre wrote:
>If any of you are familiar with mailgraph for Postfix-based mail
>servers, I''ve created something similar for Fedora DS.  fdsgraph
tails
>  
>
For anyone that uses an SNMP based grapher like cricket or cacti,
note that most of this information is also available via SNMP (and LDAP
too in the monitor entry). There''s quite a bit of potentially useful
data
in the montitor entries that does not show up in the access log also
(database activity for example).

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

David Boreham wrote:
> Chris St. Pierre wrote:
> 
>> If any of you are familiar with mailgraph for Postfix-based mail
>> servers, I''ve created something similar for Fedora DS. 
fdsgraph tails
>>  
>>
> For anyone that uses an SNMP based grapher like cricket or cacti,
> note that most of this information is also available via SNMP (and LDAP
> too in the monitor entry). There''s quite a bit of potentially
useful data
> in the montitor entries that does not show up in the access log also
> (database activity for example).
monitor entry?
what is that?
how can i access the monitor entry?

thanks
sigid
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFFRUU6qiPNNgPlDu0RAtzwAJ9iX4otIxYgf3Xd8U/DtQbgT61X/QCfQd7/
OVKj7KSidJzoh58xcBSLKCs=RGiI
-----END PGP SIGNATURE-----

>monitor entry?
>what is that?
>how can i access the monitor entry?
>  
>
http://www.redhat.com/docs/manuals/dir-server/ag/7.1/dsstats.html#1005738
http://www.redhat.com/docs/manuals/dir-server/ag/7.1/dsstats.html#1013682

also

http://www.redhat.com/docs/manuals/dir-server/ag/7.1/snmp.html#1073255