[Bug nscd/13931] New: Reload nscd cache entry even if its timeout is equal to the current time

[Bug nscd/13931] New: Reload nscd cache entry even if its timeout is equal to the current time

[siddhesh at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=13931

             Bug #: 13931
           Summary: Reload nscd cache entry even if its timeout is equal
                    to the current time
           Product: glibc
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: nscd
        AssignedTo: unassigned@sourceware.org
        ReportedBy: siddhesh@redhat.com
                CC: drepper.fsp@gmail.com
    Classification: Unclassified


With the current implementation of cache expiry, the timer for expiry
is set exactly at the timeout value of an entry. As a result, there is often a
case where the cache prune is called at exactly the timeout time, causing the
prune_cache code to think that the timeout has not occurred yet. As a result,
the entry is not invalidated or reloaded until the next prune cycle, which is
at least 15 seconds away. Fix is to mark an entry as timed out even when the
time that prune_cache is called is exactly equal to the timeout time of the
entry.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug nscd/13931] Reload nscd cache entry even if its timeout is equal to the current time

[siddhesh at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=13931

--- Comment #1 from Siddhesh Poyarekar <siddhesh at redhat dot com> 2012-03-30 11:10:25 UTC ---
Sorry, forgot to put in steps to reproduce:

1) Add an entry in dns server that expires in 15 seconds

ad.foo.net. 15 A 10.0.0.10

2) Start nscd with debug level > 2 in nscd.conf
3) getent hosts ad.foo.net

Actual Results:

Logs look like this:

Fri 30 Mar 2012 04:37:18 PM IST - 19240: handle_request: request received
(Version = 2) from PID 19256
Fri 30 Mar 2012 04:37:18 PM IST - 19240:     GETAI (ad.foo.net)
Fri 30 Mar 2012 04:37:18 PM IST - 19240: Haven't found "ad.foo.net" in hosts
cache!
Fri 30 Mar 2012 04:37:18 PM IST - 19240: add new entry "ad.foo.net" of type
GETAI for hosts to cache (first)
Fri 30 Mar 2012 04:37:33 PM IST - 19240: pruning hosts cache; time 1333105653
Fri 30 Mar 2012 04:37:33 PM IST - 19240: considering GETAI entry "ad.foo.net",
timeout 1333105653
Fri 30 Mar 2012 04:37:48 PM IST - 19240: pruning hosts cache; time 1333105668
Fri 30 Mar 2012 04:37:48 PM IST - 19240: considering GETAI entry "ad.foo.net",
timeout 1333105653
Fri 30 Mar 2012 04:37:48 PM IST - 19240: Reloading "ad.foo.net" in hosts cache!


Expected Results:

Logs should look like this:

Logs look like this:

Fri 30 Mar 2012 04:37:18 PM IST - 19240: handle_request: request received
(Version = 2) from PID 19256
Fri 30 Mar 2012 04:37:18 PM IST - 19240:     GETAI (ad.foo.net)
Fri 30 Mar 2012 04:37:18 PM IST - 19240: Haven't found "ad.foo.net" in hosts
cache!
Fri 30 Mar 2012 04:37:18 PM IST - 19240: add new entry "ad.foo.net" of type
GETAI for hosts to cache (first)
Fri 30 Mar 2012 04:37:33 PM IST - 19240: pruning hosts cache; time 1333105653
Fri 30 Mar 2012 04:37:33 PM IST - 19240: considering GETAI entry "ad.foo.net",
timeout 1333105653
Fri 30 Mar 2012 04:37:33 PM IST - 19240: Reloading "ad.foo.net" in hosts cache!

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug nscd/13931] Reload nscd cache entry even if its timeout is equal to the current time

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=13931

Florian Weimer <fweimer at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
              Flags|                            |security-

-- 
You are receiving this mail because:
You are on the CC list for the bug.