There was no apache involved.
This was caused by a custom perl script which queried gmetad's interactive
port for various metrics (nagios check script, to be percise).
Anyway, on single-CPU machine it worked OK because in that single jiffie
gmetad was able to form a response and send it back. Howe
Hi Bostjan:
On Wed, Nov 3, 2010 at 6:14 PM, Bostjan Skufca wrote:
> I was just about to ask what this error (in subject) means but then I
> discovered the solution. The same script for accessing gmetad's interactive
> port, which worked fine on another machine (identically configured gmetad),
>
After I ran out of things to try, I went digging through apache's
error logs and found this gem:
PHP Fatal error: Allowed memory size of 8388608 bytes exhausted
(tried to allocate 78 bytes) in /var/www/ganglia/ganglia.php on line
110
The (RedHat?) default of 8 MB can be changed in the main php.
James,
cool. No need to be sorry. This is actually valuable information, as
this may hit others as well.
How did you find out and where exactely is the php_value located in
the config files?
Thanks
Martin
--- James Trater <[EMAIL PROTECTED]> wrote:
> I figured it out. I had assumed that it w
I figured it out. I had assumed that it was a gmetad problem, but it
turned out to be a problem with PHP - specifically the amount of
memory that PHP is allowed to allocate. I put this in my apache config
file for ganglia:
php_value memory_limit 32M
and it works fine now. Sorry!
Jim
I did an "strace" to look at it more closely. It looks like it is
"timing out" before it is finished writing the XML. Here is a partial
view of the strace output. Things look pretty normal until the end. It
never writes the closing statement and it gives an error
message.
Assuming I'm on the right
Hi Martin.
Unfortunately,
Running gmetad in debugging mode doesn't show much more (see below.)
gmetad -d 10
--
Writing Summary data for source CCSD, metric swap_free
Writing Summary data for source CCSD, metric mem_cach
Jim, [adding the developers to CC]
good - I did not break it :-) We will need to have a closer look.
could you please run "gmetad" with "-d 1" or larger? Maybe we can track
it down.
Martin
--- James Trater <[EMAIL PROTECTED]> wrote:
> I had the same issue with 3.0.2 .
>
> -Jim
>
> On 4/18/06
I had the same issue with 3.0.2 .
-Jim
On 4/18/06, Martin Knoblauch <[EMAIL PROTECTED]> wrote:
> James,
>
> is this new with 3.0.3, or did it also happen with 3.0.2 or earlier
> versions?
>
> Thanks
> Martin
>
> --- James Trater <[EMAIL PROTECTED]> wrote:
>
> > Hi,
> >
> > I am running gmetad 3.
James,
is this new with 3.0.3, or did it also happen with 3.0.2 or earlier
versions?
Thanks
Martin
--- James Trater <[EMAIL PROTECTED]> wrote:
> Hi,
>
> I am running gmetad 3.0.3 to pull data from 3 clusters in our grid.
> Two of the cluster are around 85 nodes and the third is 320+
> nodes.
10 matches
Mail list logo
