Dear Developers,
our IT staff got up early: Saros-Build is alive and runs on two cores.
Franz
-Original Message-
From: se_p_agil-bounces~franz.zieris=fu-berlin...@lists.spline.inf.fu-berlin.de
[mailto:se_p_agil-bounces~franz.zieris=fu-berlin...@lists.spline.inf.fu-berlin.de]
On Behalf
Dear Developers,
Currently our Jenkins is not very responsive, the two CPUs are far below 10%,
whereas Gerrit is working fast and fine. I already restarted tomcat6
(/etc/init.d/tomcat6 restart), but it didn't work out. Any ideas?
Franz
--
Dear Developers,
I already read the tomcat logfiles, but didn't find anything promising.
My current guess is, that there were some flawed build directories (build
started, but did not finish) and the subsequent build tried to use the folder
again. Currently I'm running a single job and put the
Mem: 4059844k total, 4020040k used,39804k free, 197924k buffers
Swap: 2097144k total,0k used, 2097144k free, 1754424k cached
WTF ?!
Am 17.09.2012 16:25, schrieb Zieris, Franz:
> Dear Developers,
>
> I already read the tomcat logfiles, but didn't find anything promising.
>
> M
So, any ideas?
-Original Message-
From: Stefan Rossbach [mailto:srossb...@arcor.de]
Sent: Montag, 17. September 2012 16:30
To: Zieris, Franz
Cc: dpp-devel@lists.sourceforge.net
Subject: Re: [Dpp-devel] Jenkins is acting ... strange
Mem: 4059844k total, 4020040k used,39804k free,
According to 'top' command: tomcat is using up to 200% cpu time (about
100% cpu time are used in the kernel / system).
Am 17.09.2012 16:36, schrieb Zieris, Franz:
> So, any ideas?
>
> -Original Message-
> From: Stefan Rossbach [mailto:srossb...@arcor.de]
> Sent: Montag, 17. September 2012
I asked our IT staff to downgrade our VM to one CPU, in order to narrow down
the possible sources of this behavior.
I'll shutdown saros-build now.
-Original Message-
From: Stefan Rossbach [mailto:srossb...@arcor.de]
Sent: Montag, 17. September 2012 16:42
To: Zieris, Franz
Cc: dpp-devel
It seems that Findbugs is causing that issue. I have to figure out how
much heap memory is currently used by the JVM.
Am 17.09.2012 16:56, schrieb Zieris, Franz:
> I asked our IT staff to downgrade our VM to one CPU, in order to narrow down
> the possible sources of this behavior.
>
> I'll shutd
OK, saros-build now running on 1 CPU.
> I have to figure out how much heap memory is currently used by the JVM.
Please let me know, whether and how I may help.
-Original Message-
From: Stefan Rossbach [mailto:srossb...@arcor.de]
Sent: Montag, 17. September 2012 17:02
To: Zieris, Franz
There seems to be a misconfiguration with the swap file.
I am trying to allocate 3 GB in 1 MB chunks, swap memory is not used,
CPU load = 100% (0% user, 100% kernel)
Am 17.09.2012 17:11, schrieb Zieris, Franz:
> OK, saros-build now running on 1 CPU.
>
>> I have to figure out how much heap memor
Just an update,
swap space is working, but I have no clue why allocating 3 GB
tooks about 5 minutes.
Am 17.09.2012 17:31, schrieb Stefan Rossbach:
> There seems to be a misconfiguration with the swap file.
>
> I am trying to allocate 3 GB in 1 MB chunks, swap memory is not used,
>
> CPU load = 10
OK, is this a pure internal issue, or are there any chances that the
hardware/external VM configuration is responsible for this?
If the latter seems likely, could you briefly summarize your steps, so I can
file a ticket for our IT staff?
-Original Message-
From: Stefan Rossbach [mailto:
Seems the JVM is performing very bad when allocating new memory to
increase it heap size, so blowing up the
Ubuntu memory manager ... I just rerun the test, now it allocates 2.5 GB
very fast, the 0.5 GB took about a minute.
Maybe Holger can help, I do not know any internals about Unix memory
ma
Ok I read the JVM (Oracle Spec)
JVM_DEFAULT_MAX_HEAP = Math.min(0.25 * totalMemory, 1 GB)
What I do not understand, normally a JVM starts with approx 64 MB heap
memory, increasing the heap as needed.
Findbugs allocates much of memory. Since the Maxheap is set to 1GB it
seems that the Garbage C
14 matches
Mail list logo
