You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@spamassassin.apache.org by bu...@bugzilla.spamassassin.org on 2008/05/08 00:26:39 UTC
[Bug 5903] New: Possible scheduling algorithm flaw
https://issues.apache.org/SpamAssassin/show_bug.cgi?id=5903
Summary: Possible scheduling algorithm flaw
Product: Spamassassin
Version: 3.2.4
Platform: Other
OS/Version: All
Status: NEW
Severity: normal
Priority: P5
Component: spamc/spamd
AssignedTo: dev@spamassassin.apache.org
ReportedBy: cblaise@esoft.com
On my 3.2.4 system, spamd happened to start at near the PID # high threshold on
a Linux 2.4 system:
mail 31857 0.0 2.6 35140 27488 ? S 10:17 0:11 spamd child
mail 31858 0.0 2.1 34180 22528 ? S 10:17 0:01 spamd child
mail 31859 0.0 1.6 33196 16508 ? S 10:17 0:00 spamd child
mail 31860 0.0 1.5 33204 16228 ? S 10:17 0:00 spamd child
mail 31861 0.0 1.5 33212 16332 ? S 10:17 0:00 spamd child
mail 31862 0.0 1.6 33220 16420 ? S 10:17 0:00 spamd child
mail 31863 0.0 1.5 33228 16120 ? S 10:17 0:00 spamd child
root 31768 0.0 1.6 33228 17368 ? S 10:17 0:08 /usr/bin/perl -T
-w /usr/bin/spamd -d -x --max-conn-per-child=20 --socketpath=/dev/spamd
--socketgroup=mail -m8 --min-spare=8 -u mail -D
As a result, after the first child hit its limit, a new one was spawned...after
the PID set had turned over. As a result of the use of the
compute_lowest_child_pid() method in SpamdForkScaling.pm, the lowest child pid
is always the newest child at least until we reach the max PID count again.
In my case, none of the originally spawned pids are being used. Instead, the
newest is being used such as:
mail 2478 1.2 4.2 49752 43212 ? S 15:39 0:13 spamd child
Then:
mail 6308 0.6 2.6 34608 27132 ? S 16:10 0:01 spamd child
I think the correct algorithm should be to use the oldest aged child and not
the lowest pid #.
--
Configure bugmail: https://issues.apache.org/SpamAssassin/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
[Bug 5903] Possible scheduling algorithm flaw
Posted by bu...@bugzilla.spamassassin.org.
https://issues.apache.org/SpamAssassin/show_bug.cgi?id=5903
Justin Mason <jm...@jmason.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |WORKSFORME
--- Comment #1 from Justin Mason <jm...@jmason.org> 2008-05-08 01:51:53 PST ---
no, this is working as designed. The idea is simply that one of the kids is
used more heavily than others, so that it remains swapped in, in-cache, etc.
while the others can be swapped out. There's no need for that kid to
specifically be the first one started.
--
Configure bugmail: https://issues.apache.org/SpamAssassin/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.