From benno@columbia.edu Thu Oct 19 12:54:13 2000
Received: from MIT.EDU (SOUTH-STATION-ANNEX.MIT.EDU [18.72.1.2])
by rt-11.mit.edu (8.9.3/8.9.3) with SMTP id MAA03919
for <bugs@RT-11.MIT.EDU>; Thu, 19 Oct 2000 12:54:12 -0400 (EDT)
Received: from ramjet.cc.columbia.edu by MIT.EDU with SMTP
id AA15177; Thu, 19 Oct 00 12:53:59 EDT
Received: (from benno@localhost)
by ramjet.cc.columbia.edu (8.9.3/8.9.3) id MAA07661;
Thu, 19 Oct 2000 12:54:11 -0400 (EDT)
Message-Id: <200010191654.MAA07661@ramjet.cc.columbia.edu>
Date: Thu, 19 Oct 2000 12:54:11 -0400 (EDT)
From: benno@columbia.edu
Reply-To: benno@columbia.edu
To: krb5-bugs@MIT.EDU
Cc: benno@columbia.edu
Subject: kdb5_util dump loops
X-Send-Pr-Version: 3.99
System: SunOS ramjet 5.8 Generic sun4u sparc SUNW,Ultra-80
Architecture: sun4
kdb5_util dump, which we use to propagate the database nightly from our
master to our slave, has recently begun looping back to handles beginning
with 'p' when it hits handles beginning with 'r'. It continues looping until
the file reaches the maximum size supported by the OS.
There have been no changes to our configuration, other than the
normal creation and modification of handles. Examination of the principal
file at the location where the loop begins does not reveal any obvious
problems, although this can not be ruled out.
Running 'kdb5_util dump tofile' with our principal data reproducibly causes
this. Unfortunately, we do not have a way of reproducing this error
otherwise. If there is any help we can provide, or possible code segments
that might be likely candidates, please let us know.
Responsible-Changed-From-To: krb5-unassigned->tlyu
Responsible-Changed-By: tlyu
Responsible-Changed-When: Wed Jan 24 17:41:44 2001
Responsible-Changed-Why:
Mine now.
From: Tom Yu <tlyu@MIT.EDU>
To: benno@columbia.edu
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-admin/896: kdb5_util dump loops
Date: Wed, 24 Jan 2001 17:41:36 -0500 (EST)
How as this database created originally? Was it created with 1.2.1?
Can you do a "getprincs" inside kadmin.local without it looping?
---Tom
From: Benn Oshrin <benno@columbia.edu>
To: Tom Yu <tlyu@MIT.EDU>
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-admin/896: kdb5_util dump loops
Date: Thu, 08 Feb 2001 15:28:47 -0500
On 01/24/01 05:41:36 PM -0500, Tom Yu wrote:
The database was originally converted from our old Kerb IV servers using
1.1.1 and db 2.7.5. We upgraded our KDC to 1.2.1 and db 2.7.7, the
configuration we now have in production, but this combination could not
iterate the database. getprincs within kadmin.local looped as well.
Fortunately, we never toss our logs, and we were able to reconstruct a list
of every principal in the KDC. Then, using a locally built program, we
extracted (essentially a getprinc operation) each handle into a new dump
file and reloaded the database. This corrected the problem, albeit in a
roundabout manner.
-Benn-
Received: from MIT.EDU (SOUTH-STATION-ANNEX.MIT.EDU [18.72.1.2])
by rt-11.mit.edu (8.9.3/8.9.3) with SMTP id MAA03919
for <bugs@RT-11.MIT.EDU>; Thu, 19 Oct 2000 12:54:12 -0400 (EDT)
Received: from ramjet.cc.columbia.edu by MIT.EDU with SMTP
id AA15177; Thu, 19 Oct 00 12:53:59 EDT
Received: (from benno@localhost)
by ramjet.cc.columbia.edu (8.9.3/8.9.3) id MAA07661;
Thu, 19 Oct 2000 12:54:11 -0400 (EDT)
Message-Id: <200010191654.MAA07661@ramjet.cc.columbia.edu>
Date: Thu, 19 Oct 2000 12:54:11 -0400 (EDT)
From: benno@columbia.edu
Reply-To: benno@columbia.edu
To: krb5-bugs@MIT.EDU
Cc: benno@columbia.edu
Subject: kdb5_util dump loops
X-Send-Pr-Version: 3.99
Show quoted text
>Number: 896
>Category: krb5-admin
>Synopsis: kdb5_util dump loops
>Confidential: no
>Severity: serious
>Priority: high
>Responsible: tlyu
>State: open
>Class: sw-bug
>Submitter-Id: unknown
>Arrival-Date: Thu Oct 19 12:55:00 EDT 2000
>Last-Modified: Fri Sep 14 13:44:44 EDT 2001
>Originator: Benn Oshrin <benno@columbia.edu>
>Organization:
>Release: krb5-1.2.1, changes to application servers, not to KDC
>Environment:
>Category: krb5-admin
>Synopsis: kdb5_util dump loops
>Confidential: no
>Severity: serious
>Priority: high
>Responsible: tlyu
>State: open
>Class: sw-bug
>Submitter-Id: unknown
>Arrival-Date: Thu Oct 19 12:55:00 EDT 2000
>Last-Modified: Fri Sep 14 13:44:44 EDT 2001
>Originator: Benn Oshrin <benno@columbia.edu>
>Organization:
>Release: krb5-1.2.1, changes to application servers, not to KDC
>Environment:
System: SunOS ramjet 5.8 Generic sun4u sparc SUNW,Ultra-80
Architecture: sun4
Show quoted text
>Description:
kdb5_util dump, which we use to propagate the database nightly from our
master to our slave, has recently begun looping back to handles beginning
with 'p' when it hits handles beginning with 'r'. It continues looping until
the file reaches the maximum size supported by the OS.
There have been no changes to our configuration, other than the
normal creation and modification of handles. Examination of the principal
file at the location where the loop begins does not reveal any obvious
problems, although this can not be ruled out.
Show quoted text
>How-To-Repeat:
Running 'kdb5_util dump tofile' with our principal data reproducibly causes
this. Unfortunately, we do not have a way of reproducing this error
otherwise. If there is any help we can provide, or possible code segments
that might be likely candidates, please let us know.
Show quoted text
>Fix:
Show quoted text
>Audit-Trail:
Responsible-Changed-From-To: krb5-unassigned->tlyu
Responsible-Changed-By: tlyu
Responsible-Changed-When: Wed Jan 24 17:41:44 2001
Responsible-Changed-Why:
Mine now.
From: Tom Yu <tlyu@MIT.EDU>
To: benno@columbia.edu
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-admin/896: kdb5_util dump loops
Date: Wed, 24 Jan 2001 17:41:36 -0500 (EST)
Show quoted text
>>>>> "benno" == benno <benno@columbia.edu> writes:
Show quoted text
benno> kdb5_util dump, which we use to propagate the database nightly
benno> from our master to our slave, has recently begun looping back
benno> to handles beginning with 'p' when it hits handles beginning
benno> with 'r'. It continues looping until the file reaches the
benno> maximum size supported by the OS. There have been no changes
benno> to our configuration, other than the normal creation and
benno> modification of handles. Examination of the principal file at
benno> the location where the loop begins does not reveal any obvious
benno> problems, although this can not be ruled out.
benno> from our master to our slave, has recently begun looping back
benno> to handles beginning with 'p' when it hits handles beginning
benno> with 'r'. It continues looping until the file reaches the
benno> maximum size supported by the OS. There have been no changes
benno> to our configuration, other than the normal creation and
benno> modification of handles. Examination of the principal file at
benno> the location where the loop begins does not reveal any obvious
benno> problems, although this can not be ruled out.
How as this database created originally? Was it created with 1.2.1?
Can you do a "getprincs" inside kadmin.local without it looping?
---Tom
From: Benn Oshrin <benno@columbia.edu>
To: Tom Yu <tlyu@MIT.EDU>
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-admin/896: kdb5_util dump loops
Date: Thu, 08 Feb 2001 15:28:47 -0500
On 01/24/01 05:41:36 PM -0500, Tom Yu wrote:
Show quoted text
> How as this database created originally? Was it created with 1.2.1?
> Can you do a "getprincs" inside kadmin.local without it looping?
> Can you do a "getprincs" inside kadmin.local without it looping?
The database was originally converted from our old Kerb IV servers using
1.1.1 and db 2.7.5. We upgraded our KDC to 1.2.1 and db 2.7.7, the
configuration we now have in production, but this combination could not
iterate the database. getprincs within kadmin.local looped as well.
Fortunately, we never toss our logs, and we were able to reconstruct a list
of every principal in the KDC. Then, using a locally built program, we
extracted (essentially a getprinc operation) each handle into a new dump
file and reloaded the database. This corrected the problem, albeit in a
roundabout manner.
-Benn-
Show quoted text
>Unformatted: