python-ldap-dev

ReconnectLDAPObject doesn't reconnect after main failure

[Alain S. <asp...@gm...>]

When testing ReconnectLDAPObject I found a bug.
The object doesn't reconnect after a main failure !

If I shutdown the ldap server and try a request, I get a
ldap.SERVER_DOWN, this is correct. (this is what I call the main
failure)
But if I restart the server, and retry the same request (with the same
object), I get an empty answer but no error!
I'm expecting a correct answer or an error (exception)!
I have a full script that show the problem at the end.
It look the object is in an incoherent state after the main failure,
in fact in an unauthenticated state ! I thing this is a problem with
libldap or openldap, not with python code.

I thing the main probleme is here !
Look !

l=ldap.ldapobject.ReconnectLDAPObject(ldap_url.initializeUrl())
l.simple_bind_s('cn=nobody,cn=internal,dc=asxnet,dc=loc', '***********')
print 'search', l.search_s(ldap_url.dn, ldap.SCOPE_SUBTREE, "(objectClass=*)")

works and return all object anonymous can get, but

l=ldap.ldapobject.ReconnectLDAPObject(ldap_url.initializeUrl())
print 'search', l.search_s(ldap_url.dn, ldap.SCOPE_SUBTREE, "(objectClass=*)")

this work too ! And don't give any error while their is no bind !
work like if l.simple_bind_s('', '') where used just before the search !

I wrote a patch but this is only a workaround that detect the main
failure, set a flag and force a reconnect before any request if the
flag is set.

Here is the output of my test case
I use a modified python-ldap, that include the patch posted in my
previous post that enable reconnect to work with whoami_s()

-- without debuging --

OpenPKG: stop: openldap.
OpenPKG: start: openldap.
Connected
whoami dn:cn=nobody,cn=internal,dc=asxnet,dc=loc
OpenPKG: stop: openldap.
OpenPKG: start: openldap.
whoami dn:cn=nobody,cn=internal,dc=asxnet,dc=loc
reconnect ok
OpenPKG: stop: openldap.
ok: ldap.SERVER_DOWN, server is realy down
OpenPKG: start: openldap.
whoami
It look i'am connected, but like anonymous

-- with debuging --

OpenPKG: stop: openldap.
OpenPKG: start: openldap.
*** ldap://localhost:389 - ReconnectLDAPObject.set_option ((17, 3),{})
*** ldap://localhost:389 - ReconnectLDAPObject.simple_bind
(('cn=nobody,cn=internal,dc=asxnet,dc=loc',
'iMmTWz5pJ+lwY7i6M/BU61ngo1aBLyqQhRrrKbEc', None, None),{})
*** ldap://localhost:389 - ReconnectLDAPObject.result3 ((1, 1, -1),{})
Connected
*** ldap://localhost:389 - ReconnectLDAPObject.whoami_s ((None, None),{})
whoami dn:cn=nobody,cn=internal,dc=asxnet,dc=loc
OpenPKG: stop: openldap.
OpenPKG: start: openldap.
*** ldap://localhost:389 - ReconnectLDAPObject.whoami_s ((None, None),{})
*** Try 1. reconnect to ldap://localhost:389...
*** ldap://localhost:389 - ReconnectLDAPObject.set_option ((17, 3),{})
*** ldap://localhost:389 - ReconnectLDAPObject.simple_bind
(('cn=nobody,cn=internal,dc=asxnet,dc=loc',
'iMmTWz5pJ+lwY7i6M/BU61ngo1aBLyqQhRrrKbEc', None, None),{})
*** ldap://localhost:389 - ReconnectLDAPObject.result3 ((1, 1, -1),{})
*** 1. reconnect to ldap://localhost:389 successful, last operation
will be repeated
*** ldap://localhost:389 - ReconnectLDAPObject.whoami_s ((None, None),{})
whoami dn:cn=nobody,cn=internal,dc=asxnet,dc=loc
reconnect ok
OpenPKG: stop: openldap.
*** ldap://localhost:389 - ReconnectLDAPObject.whoami_s ((None, None),{})
*** Try 1. reconnect to ldap://localhost:389...
*** ldap://localhost:389 - ReconnectLDAPObject.set_option ((17, 3),{})
*** ldap://localhost:389 - ReconnectLDAPObject.simple_bind
(('cn=nobody,cn=internal,dc=asxnet,dc=loc',
'iMmTWz5pJ+lwY7i6M/BU61ngo1aBLyqQhRrrKbEc', None, None),{})
*** 1. reconnect to ldap://localhost:389 failed
ok: ldap.SERVER_DOWN, server is realy down
OpenPKG: start: openldap.
*** ldap://localhost:389 - ReconnectLDAPObject.whoami_s ((None, None),{})
whoami
It look i'am connected, but like anonymous

---- and finaly my test case ----

import sys, os, time
import ldap, ldapurl

host='localhost'
port=389
who='cn=nobody,cn=internal,dc=asxnet,dc=loc'
cred='iMmTWz5pJ+lwY7i6M/BU61ngo1aBLyqQhRrrKbEc'
dn='dc=asxnet,dc=loc'

def ldap_service(action):
    os.system('/kolab/bin/openpkg rc openldap %s' % action)
    if action.endswith('start'):
        time.sleep(1)

def check_connection():
    whoami=l.whoami_s()
    print 'whoami', whoami
    # this search dont give any result as anonymous, but well if
loggged as nobody
    #result=l.search_s(ldap_url.dn, ldap.SCOPE_SUBTREE,
"(member=cn=domain.maintainer
mydomain.loc,cn=internal,dc=asxnet,dc=loc)")
    #print 'search', result

ldap_url=ldapurl.LDAPUrl('ldap://%s:%d/%s' % (host, port, dn))
ldap_url.applyDefaults({
    'who': who,
    'cred' : cred, })

# to be sure the server is up
ldap_service('stop')
ldap_service('start')

l=ldap.ldapobject.ReconnectLDAPObject(ldap_url.initializeUrl(), 1)
# l=ldap.ldapobject.LDAPObject(ldap_url.initializeUrl())
l.simple_bind_s(ldap_url.who, ldap_url.cred)
print 'Connected'

check_connection()

ldap_service('stop')
ldap_service('start')

try:
    check_connection()
except ldap.SERVER_DOWN:
    print "Error: ldap.SERVER_DOWN !"
else:
    print "reconnect ok"

ldap_service('stop')

try:
    check_connection()
except ldap.SERVER_DOWN:
    print "ok: ldap.SERVER_DOWN, server is realy down"

ldap_service('start')

check_connection()

print "It look i'am connected, but like anonymous"








import sys, os, time
import ldap, ldapurl

host='localhost'
port=389
who='cn=nobody,cn=internal,dc=asxnet,dc=loc'
cred='iMmTWz5pJ+lwY7i6M/BU61ngo1aBLyqQhRrrKbEc'
dn='dc=asxnet,dc=loc'

def ldap_service(action):
    os.system('/kolab/bin/openpkg rc openldap %s' % action)
    if action.endswith('start'):
        time.sleep(1)

def check_connection():
    #print 'search', l.search_s(ldap_url.dn, ldap.SCOPE_SUBTREE,
"(member=cn=domain.maintainer
mydomain.loc,cn=internal,dc=asxnet,dc=loc)")
    print 'whoami', l.whoami_s()

ldap_url=ldapurl.LDAPUrl('ldap://%s:%d/%s' % (host, port, dn))
ldap_url.applyDefaults({
    'who': who,
    'cred' : cred, })

ldap_service('stop')
ldap_service('start')

l=ldap.ldapobject.ReconnectLDAPObject(ldap_url.initializeUrl())
# l=ldap.ldapobject.LDAPObject(ldap_url.initializeUrl())
l.simple_bind_s(ldap_url.who, ldap_url.cred)
print 'Connected'

check_connection()

ldap_service('stop')
ldap_service('start')

try:
    check_connection()
except ldap.SERVER_DOWN:
    print "Error: ldap.SERVER_DOWN !"

ldap_service('stop')

try:
    check_connection()
except ldap.SERVER_DOWN:
    print "Ok: ldap.SERVER_DOWN"

ldap_service('start')

check_connection()


ANY Comments ?


--
Alain Spineux
aspineux gmail com
May the sources be with you

Re: ReconnectLDAPObject doesn't reconnect after main failure

[<mi...@st...>]

Alain Spineux wrote:
> When testing ReconnectLDAPObject I found a bug.
> The object doesn't reconnect after a main failure !
> [..]
> I thing the main probleme is here !
> Look !

Yes, I can reproduce your observation. Thanks for pointing it out.

> l=ldap.ldapobject.ReconnectLDAPObject(ldap_url.initializeUrl())
> print 'search', l.search_s(ldap_url.dn, ldap.SCOPE_SUBTREE, "(objectClass=*)")
> 
> this work too ! And don't give any error while their is no bind !
> work like if l.simple_bind_s('', '') where used just before the search !

Yes, this works as intended in LDAPv3. In opposite to LDAPv2 you MAY
send a LDAP request without prior bind request.

> I wrote a patch but this is only a workaround that detect the main
> failure, set a flag and force a reconnect before any request if the
> flag is set.

Where's the patch? Can't figure out why it does not send the formerly
sent bind request. It has all the data around. And if you take the
server down it will re-send the bind request.

Ciao, Michael.

Re: ReconnectLDAPObject doesn't reconnect after main failure

[Alain S. <asp...@gm...>]

On 1/26/07, Michael Str=F6der <mi...@st...> wrote:
> Alain Spineux wrote:
> > When testing ReconnectLDAPObject I found a bug.
> > The object doesn't reconnect after a main failure !
> > [..]
> > I thing the main probleme is here !
> > Look !
>
> Yes, I can reproduce your observation. Thanks for pointing it out.

Don't forget we are on the same boat Captain ! :-)
>
> > l=3Dldap.ldapobject.ReconnectLDAPObject(ldap_url.initializeUrl())
> > print 'search', l.search_s(ldap_url.dn, ldap.SCOPE_SUBTREE, "(objectCla=
ss=3D*)")
> >
> > this work too ! And don't give any error while their is no bind !
> > work like if l.simple_bind_s('', '') where used just before the search =
!
>
> Yes, this works as intended in LDAPv3. In opposite to LDAPv2 you MAY
> send a LDAP request without prior bind request.
>
> > I wrote a patch but this is only a workaround that detect the main
> > failure, set a flag and force a reconnect before any request if the
> > flag is set.
>
> Where's the patch?
This is still only a workin "draft" I need to reread it's logic,
anyway I add it at the end
> Can't figure out why it does not send the formerly
> sent bind request. It has all the data around. And if you take the
> server down it will re-send the bind request.

The trick is it doesn't try to reconnect AFTER the main failure (or
BEFORE the next statement) !
Because the first try (of the next ldap statement) doesn't generate
any error, it doesn't try to reconnect and then give the answer as is
! But this is the answer it got like an unauthenticated user !

Look, this is the tcpdump of the last statement !
00:14:21.109514 IP 127.0.0.1.36250 > 127.0.0.1.ldap: S
1795648490:1795648490(0) win 00:14:21.110352 IP 127.0.0.1.ldap >
127.0.0.1.36250: S 1793421325:1793421325(0) ack 1795648491 win 32768
<mss 16396,sackOK,timestamp 516494160 516494159,nop,wscale 7>
00:14:21.110430 IP 127.0.0.1.36250 > 127.0.0.1.ldap: . ack 1 win 257
<nop,nop,timestamp 516494160 516494160>

You can identify the well known 3-way handshaking ! This is a new connectio=
n !
I thing libldap open a new connection, but libldap dont store anything
about the credential and then cannot authenticate with the good user !
I will try to reproduce this in C with libc

Here is the patch (including my whoami_s() )

*** ldapobject.py.disconnet     Fri Jan 26 22:22:57 2007
--- ldapobject.py.orig  Fri Jan 26 13:42:59 2007
***************
*** 678,684 ****
      self._retry_delay =3D retry_delay
      self._start_tls =3D 0
      self._reconnects_done =3D 0L
-     self._disconnected=3DTrue

    def __getstate__(self):
      """return data representation for pickled object"""
--- 678,683 ----
***************
*** 723,730 ****
            self.start_tls_s()
          # Repeat last simple or SASL bind
          self._apply_last_bind()
!       except Exception, e:
!         self._disconnected=3DTrue
          if __debug__ and self._trace_level>=3D1:
            self._trace_file.write('*** %d. reconnect to %s failed\n' % (
              self._retry_max-reconnect_counter+1,uri
--- 722,728 ----
            self.start_tls_s()
          # Repeat last simple or SASL bind
          self._apply_last_bind()
!       except:
          if __debug__ and self._trace_level>=3D1:
            self._trace_file.write('*** %d. reconnect to %s failed\n' % (
              self._retry_max-reconnect_counter+1,uri
***************
*** 736,742 ****
            self._trace_file.write('=3D> delay %s...\n' % (self._retry_dela=
y))
          time.sleep(self._retry_delay)
        else:
-         self._disconnected=3DFalse
          if __debug__ and self._trace_level>=3D1:
            self._trace_file.write('*** %d. reconnect to %s
successful, last operation will be repeated\n' % (
              self._retry_max-reconnect_counter+1,uri
--- 734,739 ----
***************
*** 745,762 ****
          break

    def _apply_method_s(self,func,*args,**kwargs):
!     if self._disconnected:
        self.reconnect(self._uri)
        # Re-try last operation
        return func(self,*args,**kwargs)
-     else:
-         try:
-           return func(self,*args,**kwargs)
-         except ldap.SERVER_DOWN:
-           # Reconnect
-           self.reconnect(self._uri)
-           # Re-try last operation
-           return func(self,*args,**kwargs)

    def set_option(self,option,invalue):
      self._options[option] =3D invalue
--- 742,754 ----
          break

    def _apply_method_s(self,func,*args,**kwargs):
!     try:
!       return func(self,*args,**kwargs)
!     except ldap.SERVER_DOWN:
!       # Reconnect
        self.reconnect(self._uri)
        # Re-try last operation
        return func(self,*args,**kwargs)

    def set_option(self,option,invalue):
      self._options[option] =3D invalue
***************
*** 764,772 ****

    def simple_bind_s(self,*args,**kwargs):
      self._last_bind =3D (self.simple_bind_s,args,kwargs)
!     result=3DSimpleLDAPObject.simple_bind_s(self,*args,**kwargs)
!     self._disconnected=3DFalse
!     return result

    def start_tls_s(self):
      res =3D SimpleLDAPObject.start_tls_s(self)
--- 756,762 ----

    def simple_bind_s(self,*args,**kwargs):
      self._last_bind =3D (self.simple_bind_s,args,kwargs)
!     return SimpleLDAPObject.simple_bind_s(self,*args,**kwargs)

    def start_tls_s(self):
      res =3D SimpleLDAPObject.start_tls_s(self)
***************
*** 778,786 ****
      sasl_interactive_bind_s(who, auth) -> None
      """
      self._last_bind =3D (self.sasl_interactive_bind_s,args,kwargs)
!     result=3DSimpleLDAPObject.sasl_interactive_bind_s(self,*args,**kwargs=
)
!     self._disconnected=3DFalse
!     return result

    def add_ext_s(self,*args,**kwargs):
      return self._apply_method_s(SimpleLDAPObject.add_ext_s,*args,**kwargs=
)
--- 768,774 ----
      sasl_interactive_bind_s(who, auth) -> None
      """
      self._last_bind =3D (self.sasl_interactive_bind_s,args,kwargs)
!     return SimpleLDAPObject.sasl_interactive_bind_s(self,*args,**kwargs)

    def add_ext_s(self,*args,**kwargs):
      return self._apply_method_s(SimpleLDAPObject.add_ext_s,*args,**kwargs=
)
***************
*** 800,808 ****
    def search_ext_s(self,*args,**kwargs):
      return self._apply_method_s(SimpleLDAPObject.search_ext_s,*args,**kwa=
rgs)

-   def whoami_s(self,*args,**kwargs):
-     return self._apply_method_s(SimpleLDAPObject.whoami_s,*args,**kwargs)
-

  class SmartLDAPObject(ReconnectLDAPObject):
    """
--- 788,793 ----



--
Alain Spineux
aspineux gmail com
May the sources be with you

Re: ReconnectLDAPObject doesn't reconnect after main failure

[<mi...@st...>]

Alain Spineux wrote:
>>
>> Where's the patch?
> This is still only a workin "draft" I need to reread it's logic,
> anyway I add it at the end

Got your point now.
But still I don't find an elegant solution without having to handle a
'disconnected' flag in several methods.

> I thing libldap open a new connection, but libldap dont store anything
> about the credential and then cannot authenticate with the good user !
> I will try to reproduce this in C with libc

Yes, it's a bug in ReconnectLDAPObject. It doesn't have anything to do
with libldap.

Ciao, Michael.

Re: ReconnectLDAPObject doesn't reconnect after main failure

[Alain S. <asp...@gm...>]

Yes I got it !

The problem is the failed statement has initiated a new connection
(when the server is down)
calling initialize (that doesn't fail) and bind that fail but is in a
try: except:  block in ReconnectLDAPObject.reconnect() !
Then the server restart and the next statement get advantages of the
work done initialize


# first stop the service
ldap_service('stop')
# initiate a connection
l=3Dldap.initialize(ldap_url.initializeUrl()) # I dont use ReconnectLDAPObj=
ect !
try:
    l.simple_bind_s(ldap_url.who, ldap_url.cred)
except ldap.SERVER_DOWN:
    pass # bind fail like in ReconnectLDAPObject.reconnect()
ldap_service('start')
check_connection() # that return an empty string


The probleme is initialize don't initialize any connection !
from  "man 3 ldap" :

       The  basic  interaction  is  as  follows.  A session handle is creat=
ed
       using ldap_initialize(3) and set the protocol version to 3 by  calli=
ng
       ldap_set_option(3).  The underlying session is established first ope=
r-
       ation is issued.  This would generally be a Start TLS or  Bind  oper=
a-
       tion.

The problem is that the first operation after Initialize will never
put the connection into failing
state when server is down!
The problem is probably into libldap into the "auto connect at first
operation" like
described into man 3 ldap

This let us do something like

l=3Dldap.initialize(ldap_url.initializeUrl()) # I dont use ReconnectLDAPObj=
ect !

while True:
    try:
        l.simple_bind_s(ldap_url.who, ldap_url.cred)
    except ldap.SERVER_DOWN:
        time.sleep(1)
    else:
        break
print 'Connected'

I don't know if this is a feature or a bug into libldap !

Then the logic of my patch look to be the good one. I still need to
verify it once more.

I go to sleep now


On 1/26/07, Michael Str=F6der <mi...@st...> wrote:
> Alain Spineux wrote:
> > When testing ReconnectLDAPObject I found a bug.
> > The object doesn't reconnect after a main failure !
> > [..]
> > I thing the main probleme is here !
> > Look !
>
> Yes, I can reproduce your observation. Thanks for pointing it out.
>
> > l=3Dldap.ldapobject.ReconnectLDAPObject(ldap_url.initializeUrl())
> > print 'search', l.search_s(ldap_url.dn, ldap.SCOPE_SUBTREE, "(objectCla=
ss=3D*)")
> >
> > this work too ! And don't give any error while their is no bind !
> > work like if l.simple_bind_s('', '') where used just before the search =
!
>
> Yes, this works as intended in LDAPv3. In opposite to LDAPv2 you MAY
> send a LDAP request without prior bind request.
>
> > I wrote a patch but this is only a workaround that detect the main
> > failure, set a flag and force a reconnect before any request if the
> > flag is set.
>
> Where's the patch? Can't figure out why it does not send the formerly
> sent bind request. It has all the data around. And if you take the
> server down it will re-send the bind request.
>
> Ciao, Michael.
>


--=20
--
Alain Spineux
aspineux gmail com
May the sources be with you

Re: ReconnectLDAPObject doesn't reconnect after main failure

[<mi...@st...>]

Alain Spineux wrote:
> 
> The problem is the failed statement has initiated a new connection
> (when the server is down)
> calling initialize (that doesn't fail) and bind that fail but is in a
> try: except:  block in ReconnectLDAPObject.reconnect() !
> Then the server restart and the next statement get advantages of the
> work done initialize

Yes. Inspired by you mentioning libldap I remembered that unbind_s()
should be called when handling ldap.SERVER_DOWN exception to completely
drop the whole connection context.

Please test the version I've committed some minutes ago.
Thanks again.

Ciao, Michael.

Re: ReconnectLDAPObject doesn't reconnect after main failure

[Alain S. <asp...@gm...>]

Yes, we did it !

My test case is working with your CVS !
I put it at the end.

And what about the idea to put the

try:
except ldap.SERVER_DOWN,e:
   SimpleLDAPObject.unbind_s(self)

in the function _ldap_call too ?
This will correct also SimpleLDAPObject ?


import sys, os, time
import ldap, ldapurl

host=3D'localhost'
port=3D389
who=3D'cn=3Dnobody,cn=3Dinternal,dc=3Dasxnet,dc=3Dloc'
cred=3D'iMmTWz5pJ+lwY7i6M/BU61ngo1aBLyqQhRrrKbEc'
dn=3D'dc=3Dasxnet,dc=3Dloc'

def ldap_service(action):
    os.system('/kolab/bin/openpkg rc openldap %s' % action)
    if action.endswith('start'):
        time.sleep(1)

def check_connection():
    whoami=3Dl.whoami_s()
    print 'whoami', whoami
    # this search dont give any result as anonymous, but well if
loggged as nobody
    #result=3Dl.search_s(ldap_url.dn, ldap.SCOPE_SUBTREE,
"(member=3Dcn=3Ddomain.maintainer
mydomain.loc,cn=3Dinternal,dc=3Dasxnet,dc=3Dloc)")
    #print 'search', result

ldap_url=3Dldapurl.LDAPUrl('ldap://%s:%d/%s' % (host, port, dn))
ldap_url.applyDefaults({
    'who': who,
    'cred' : cred, })

# to be sure the server is up
ldap_service('stop')
ldap_service('start')

l=3Dldap.ldapobject.ReconnectLDAPObject(ldap_url.initializeUrl(),1) # I
dont use ReconnectLDAPObject !
l.simple_bind_s(ldap_url.who, ldap_url.cred)
check_connection()

print 'Wait 120s'
time.sleep(120)
check_connection()

print 'restart service'
ldap_service('stop')
ldap_service('start')
check_connection()

print 'stop service'
ldap_service('stop')
try:
    check_connection()
except ldap.SERVER_DOWN:
    print 'check failed, OK'

print 'restart service'
ldap_service('start')
check_connection()




On 1/27/07, Michael Str=F6der <mi...@st...> wrote:
> Alain Spineux wrote:
> >
> > The problem is the failed statement has initiated a new connection
> > (when the server is down)
> > calling initialize (that doesn't fail) and bind that fail but is in a
> > try: except:  block in ReconnectLDAPObject.reconnect() !
> > Then the server restart and the next statement get advantages of the
> > work done initialize
>
> Yes. Inspired by you mentioning libldap I remembered that unbind_s()
> should be called when handling ldap.SERVER_DOWN exception to completely
> drop the whole connection context.
>
> Please test the version I've committed some minutes ago.
> Thanks again.
>
> Ciao, Michael.
>


--=20
--
Alain Spineux
aspineux gmail com
May the sources be with you

Re: ReconnectLDAPObject doesn't reconnect after main failure

[<mi...@st...>]

On 12:59:58 pm 2007-01-27 "Alain Spineux" <asp...@gm...> wrote:
>
> And what about the idea to put the
>
> try:
> except ldap.SERVER_DOWN,e:
>    SimpleLDAPObject.unbind_s(self)
>
> in the function _ldap_call too ?
> This will correct also SimpleLDAPObject ?

There is nothing wrong with SimpleLDAPObject. The application is supposed
to catch and handle ldap.SERVER_DOWN. Additionally SimpleLDAPObject and
ReconnectLDAPObject shall behave in the same way. And IMO they do now.

Ciao, Michael.

Re: ReconnectLDAPObject doesn't reconnect after main failure

[Alain S. <asp...@gm...>]

This is a good point of view!
Anyway I try my _LAST_ arguments.

Python is a language made to help and facilitate the developer work :
- the developer dont need to test the result of any function, python
(or the library) raise an exception if something is wrong.
- the developer dont need to worry about the memory allocation, the
garbage collector do it for him.
- the developer dont need to close a file, the system do it for him
when the object is released
- the developer don't need to worry for long living LDAP connection,
ReconnectLDAPObject  auto reconnect automatically for him :-)
- python-ldap is also thread safe .....

Then why does the developer have to encapsulate any ldap statement into a

try:
except ldap.SERVER_DOWN,e:
    SimpleLDAPObject.unbind_s(self)

whereas=09the library can do it for him

Anyway we made a good job !

Tanks for your support.

On 1/27/07, "Michael Str=F6der" <mi...@st...> wrote:
> On 12:59:58 pm 2007-01-27 "Alain Spineux" <asp...@gm...> wrote:
> >
> > And what about the idea to put the
> >
> > try:
> > except ldap.SERVER_DOWN,e:
> >    SimpleLDAPObject.unbind_s(self)
> >
> > in the function _ldap_call too ?
> > This will correct also SimpleLDAPObject ?
>
> There is nothing wrong with SimpleLDAPObject. The application is supposed
> to catch and handle ldap.SERVER_DOWN. Additionally SimpleLDAPObject and
> ReconnectLDAPObject shall behave in the same way. And IMO they do now.
>
> Ciao, Michael.
>
>


--=20
--
Alain Spineux
aspineux gmail com
May the sources be with you

Re: ReconnectLDAPObject doesn't reconnect after main failure

[<mi...@st...>]

Alain Spineux wrote:
> 
> Python is a language made to help and facilitate the developer work :
> - the developer dont need to test the result of any function, python
> (or the library) raise an exception if something is wrong.
> - the developer dont need to worry about the memory allocation, the
> garbage collector do it for him.
> - the developer dont need to close a file, the system do it for him
> when the object is released
> - the developer don't need to worry for long living LDAP connection,
> ReconnectLDAPObject  auto reconnect automatically for him :-)
> - python-ldap is also thread safe .....

Your list rather arguments for adding SimpleLDAPObject.__del__() if not
implemented yet (I have to look at it).

> Then why does the developer have to encapsulate any ldap statement into a
> 
> try:
> except ldap.SERVER_DOWN,e:
>    SimpleLDAPObject.unbind_s(self)

Because in case of ldap.SERVER_DOWN the application might have some
means for fail-over, smart error handling or similar. python-ldap is
rather low-level. I expect application developers to wrap something
around SimpleLDAPObject which better suits there needs.

And I'd like to avoid breaking existing code. Calling unbind_s() twice
raises
ldap.LDAPError: LDAP connection invalid

Do you have an overview how other database modules handle such cases?

Ciao, Michael.

Re: ReconnectLDAPObject doesn't reconnect after main failure

[Alain S. <asp...@gm...>]

Convinced :-)

Best regards.

On 1/29/07, Michael Str=F6der <mi...@st...> wrote:
> Alain Spineux wrote:
> >
> > Python is a language made to help and facilitate the developer work :
> > - the developer dont need to test the result of any function, python
> > (or the library) raise an exception if something is wrong.
> > - the developer dont need to worry about the memory allocation, the
> > garbage collector do it for him.
> > - the developer dont need to close a file, the system do it for him
> > when the object is released
> > - the developer don't need to worry for long living LDAP connection,
> > ReconnectLDAPObject  auto reconnect automatically for him :-)
> > - python-ldap is also thread safe .....
>
> Your list rather arguments for adding SimpleLDAPObject.__del__() if not
> implemented yet (I have to look at it).
>
> > Then why does the developer have to encapsulate any ldap statement into=
 a
> >
> > try:
> > except ldap.SERVER_DOWN,e:
> >    SimpleLDAPObject.unbind_s(self)
>
> Because in case of ldap.SERVER_DOWN the application might have some
> means for fail-over, smart error handling or similar. python-ldap is
> rather low-level. I expect application developers to wrap something
> around SimpleLDAPObject which better suits there needs.
>
> And I'd like to avoid breaking existing code. Calling unbind_s() twice
> raises
> ldap.LDAPError: LDAP connection invalid
>
> Do you have an overview how other database modules handle such cases?
>
> Ciao, Michael.
>


--=20
--
Alain Spineux
aspineux gmail com
May the sources be with you