No, 2.5.7 is exactly what its version number suggests: a small delta
from 2.5.6.
I'm not sure you all got the announcement I sent about it originally,
but the only new thing is the cleanup thread in gmetad, more accurate
TN calculations in gmetad's output, and a small patch from Rick Warner.
--(from Sept 29th)--
Gmetad 2.5.x never had a cleaup thread, which caused a type of memory
leak if there were lots of gmetrics from monitored clusters. This
version 2.5.7 fixes this problem: a gmetad cleanup thread trims the
metric hash tables of metrics whose DMAX (delete time) has been
reached.
...
The tarball is available here:
http://heron.sdsc.edu/ganglia-software/ganglia-monitor-core-2.5.7.tar.gz
--
-Federico
On Oct 7, 2004, at 2:37 PM, Adesanya, Adeyemi wrote:
Hi.
I'm swamped at the moment but I'd like to try 2.5.7 ASAP. Does it
include UDP unicasts and some of the other improvements you made to
2.5.x ????
----
Yemi
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On
Behalf Of Matt Massie
Sent: Thursday, October 07, 2004 2:10 PM
To: Ganglia Developers
Subject: [Ganglia-developers] new 3.0.0 snapshot 2.5.7 feedback
guys-
have any of you tried the new 2.5.7 snapshot that federico
announced? if i don't hear any screaming from you guys, we're
going to release it sometime at the very end of this week (or
the beginning of next week).
also
i just put out a new snapshot of 3.0.0 at
http://matt-massie.com/ganglia/.
i had to completely rewrite the memory-pool code (in
./lib/pool.c) to squash the bug. the new pool code is
cleaner as well as missing memory leaks in certain boundary
conditions. i've been running the same test i did before
(where i have clients slam gangliad for data). for about 1/2
hour and hit it over 250,000 times without any leaks at all.
over the last few days, valgrind told me that there might be
a leak in the pool functions but there were also many other
suspicious memory areas as well (i wasted time tracking down
all these areas to find they were okay). i ignored the
warnings in part because i wrote a very simple pool test (in
./tests) that didn't show any problems so i ignored those
messages about the pool code.
what i learned is that the old memory pool library wouldn't
let you create a subpool, allocate a new structure from the
subpool, and then set an element in the structure to the
subpool. this trick allow a single pool_free of the subpool
to also free the allocated structure.
the new library doesn't have that limitation....
sub_pool = pool_new( parent_pool );
thang = pool_alloc( sub_pool, sizeof( struct thang ));
thang->pool = sub_pool;
/* use the thang until i don't need it anymore * pool_free(
thang->pool );
/* will free all memory for the thang */
this new snapshot also outputs data in valid xhtml. you
might take a look at the xhtml source. i'll be sending out
more on this soon.
hope you're all enjoying the day of thor.
-matt
--
PGP fingerprint 'A7C2 3C2F 8445 AD3C 135E F40B 242A 5984 ACBC 91D3'
They that can give up essential liberty to obtain a little
temporary safety deserve neither liberty nor safety.
--Benjamin Franklin, Historical Review of Pennsylvania, 1759
-------------------------------------------------------
This SF.net email is sponsored by: IT Product Guide on
ITManagersJournal
Use IT products in your business? Tell us what you think of them. Give
us
Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out
more
http://productguide.itmanagersjournal.com/guidepromo.tmpl
_______________________________________________
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers
Federico
Rocks Cluster Group, San Diego Supercomputer Center, CA
