I've gotten four error messages from my bot today, several relating to out of memory errors or segfaults. This bot has run without error for months, and never messes directly with memory, using the built-in Java stuff for file handling. Is something wrong with willow?
Messages in order: 14:00 UTC: Segmentation Fault - core dumped
16:00 UTC: ld.so.1: sh: fatal: mmap anon failed: Resource temporarily unavailable
Followed by: ld.so.1: uname: fatal: /lib/libc.so.1: mmap failed: Resource temporarily unavailable error: JSV stderr: Out of memory! Unable to run job: JSV stderr: Out of memory! JSV stderr is - Out of memory!. Exiting
18:00 UTC apparently ran just fine, no errors reported...
20:00 UTC: /opt/local/bin/cronsub[56]: 8239 Segmentation Fault(coredump)
---- User:Hersfold hersfoldwiki@gmail.com
Sent from my Windows Phone
On Mon, May 14, 2012 at 4:12 PM, Hersfold Wikipedia hersfoldwiki@gmail.com wrote:
I've gotten four error messages from my bot today, several relating to out of memory errors or segfaults. This bot has run without error for months, and never messes directly with memory, using the built-in Java stuff for file handling. Is something wrong with willow?
I have gotten the same errors. The "problem" is that someone is running lots of processes. The sysadmins will be able to diagnose it.
- Carl
Hello, At Monday 14 May 2012 22:51:40 DaB. wrote:
Is something wrong with willow?
yes and no: The server is just overloaded (and out of memory). Two things will fix this on short-term: The adding of new (linux-)boxes somewhen next week and the disabling of the interwiki-bots tomorrow.
Sincerely, DaB.
Last week a few Mono-powered bots have been randomly crashing as well.
Stacktrace started with: mmap(...PROT_NONE...) failed
Which indicates out of memory, indeed.
-- Krinkle
toolserver-l@lists.wikimedia.org