Marcelo,
That is my code, and you may be right. I cannot tell what is your new
logic. Can you send me a diff to the original (2.5.7) code?
Federico
On Jan 2, 2005, at 2:38 PM, Marcelo Veiga Neves wrote:
hi,
i think i found a bug in gmetad (latest version 2.5.7).
write_root_summary() writes data to RRD forever! and gmetad never
clean up metric summaries (root.metric_summary). i think that if a
metric was not modified (by do_root_summary()), then it does not have
to be written to RRD.
I do it in gmetad/gmetad.c:
static int
delete_metric(datum_t *key, datum_t *val
, void *arg)
{
Metric_t *metric = (Metric_t*)val->data;
struct cleanup_arg *cleanup = (struct cleanup_arg*)arg;
if(metric->num == 0) {
cleanup->key=key;
return 1;
}
return 0;
}
static int
cleanup_root_summary(hash_t *metric_summary)
{
datum_t *rv;
struct cleanup_arg clean;
clean.key=0;
while (hash_walkfrom(metric_summary, clean.hashval, delete_metric,
(void*)&clean)) {
if (clean.key) {
clean.hashval = hashval(clean.key, metric_summary);
rv=hash_delete(clean.key, metric_summary);
if (rv) datum_free(rv);
clean.key=0;
}
}
}
/* Meta data */
for(;;)
{
/* Do at a random interval between 10 and 30 sec. */
sleep_time = 10 + ((30-10)*1.0) * rand()/(RAND_MAX + 1.0);
sleep(sleep_time);
/* Flush the old values */
hash_foreach(root.metric_summary, zero_out_summary, NULL);
root.hosts_up = 0;
root.hosts_down = 0;
/* Sum the new values */
hash_foreach(root.authority, do_root_summary, NULL );
/* Delete dead metrics */
cleanup_root_summary(root.metric_summary);
/* Save them to RRD */
hash_foreach(root.metric_summary, write_root_summary, NULL);
}
Am I Right?
Marcelo Veiga Neves
http://www.inf.ufsm.br/~veiga
--
-- "talk is cheap, show me the code"
-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Ganglia-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ganglia-developers
Federico
Rocks Cluster Group, San Diego Supercomputer Center, CA
