This is using the latest version (updated about 40 minutes ago).
I am noticing something disturbing, though. The mute Linux node (running
2.4.1) that I do all my Linux cluster data collecting on is dying after a
few minutes (5-10 minutes, I'm timing it now).
mcast_listen_thread() got internal hash 0x809f080
mcast_listen_thread() built metricdata struct
mcast_listen_thread() attempting to hash_insert_data
mcast_listen_thread() inserted data into 0x809f080
mcast_listen_thread() got a 8 byte multicast message
mcast_listen_thread() received key 11
mcast_listen_thread() received metric data cpu_nice
1026 pre_process_node() remote_ip=10.10.9.40
pre_process_node() HOSTNAME =SOME_MEMBER.SOME_CLUSTER
pre_process_node() TIMESTAMP=1031594864
pre_process_node() HASHP =0x81343a8
pre_process_node() USER_HASHP=0x81343b8
pre_process_node() returning the ganglia internal hash pointer 0x81343a8
mcast_listen_thread() got internal hash 0x81343a8
mcast_listen_thread() built metricdata struct
mcast_listen_thread() attempting to hash_insert_data
mcast_listen_thread() inserted data into 0x81343a8
Broken pipe
It's tracking about 250 nodes, no user-defined metrics (in other words, the
metric hash it's initializing should still be adequate for this...).
gmetad seems to be complaining about dead sources an awful lot, even though
it's getting updates at least once every 60 seconds from both the new and
old data sources. Don't know if that's cause for concern.
Periodically, I get this:
RRD_update: illegal attempt to update using time 1031595909 when last
update time is 1031595909 (minimum one second step)
RRD_update: expected 1 data source readings (got 2) from N:263.000000:143:...
save_to_rrd() couldn't parse the XML and data to RRD for [SOME_CLUSTER]
Anyway, I'll continue to prod this thing with my foot to see if I get any
hints.
matt massie wrote:
you guys need to update to the latest CVS version of the web frontend in
order to use the new gmetad.
give it a try and let me know what you find please.
-matt
Today, Preston Smith wrote forth saying...
On Mon, Sep 09, 2002 at 10:35:40AM -0700, Steven Wagner ([EMAIL PROTECTED])
wrote:
Of course there are screenloads of what appear to be real updates. But I'm
not seeing the graphs update on the frontend, which admittedly is running
on a hacked-up 0.1.0 codebase. My updated incarnation is having trouble
rendering graphs. Maybe the RRD layout has changed between versions?
When I run basic vanilla "rrdtool dump," though, I don't see any new
values. But the files *are* being written to (the timestamps are all
updating)... except for the summaries...
I see the same thing, I just updated all of my cluster to 2.5.0.
The frontend's showing values, which seem accurate, but no graphs.
