Ivan B.,
Good news, thank you!
2020-07-27 10:28 GMT+03:00, Ivan Bessonov :
> Hi Ivan P.,
>
> I configured it for both PDS (Indexing) and PDS 4 (was asked by Nikita
> Tolstunov). It totally worked, not a single 137 since then.
> Occasional 130 will be fixed in [1], it has a different problem behin
Hi Ivan P.,
I configured it for both PDS (Indexing) and PDS 4 (was asked by Nikita
Tolstunov). It totally worked, not a single 137 since then.
Occasional 130 will be fixed in [1], it has a different problem behind it.
Now I'm trying to find someone who knows TC configuration better and
will be ab
Ivan B.,
I noticed that you were able to configure environment variables for
PDS (Indexing). Do field experiments show that the suggested approach
fixes the problem?
Interesting stuff with jemalloc. It might be useful to file a ticket.
2020-07-23 16:07 GMT+03:00, Ivan Daschinsky :
>>
>> About "j
>
> About "jemalloc" - it's also an option, but it also requires reconfiguring
> suites on
> TC, maybe in a more complicated way. It requires additional installation,
> right?
> Can we stick to the solution that I already tested or should we update TC
> agents? :)
Yes, if you want to use jemalloc
>
> glibc allocator uses arenas for minimize contention between threads
I understand it the same way. I did testing with running of Indexing suite
locally
and periodically executing "pmap ", it showed that the number of 64mb
arenas grows constantly and never shrinks. By the middle of the suite th
AFAIK, glibc allocator uses arenas for minimize contention between threads
when they trying to access
or free preallocated bit of memory. But seems that we
use -XX:+AlwaysPreTouch, so heap is allocated
and committed at start time. We allocate memory for durable memory in one
thread.
So I think ther
Hello Ivan,
It feels like the problem is more about new starting threads rather than the
allocation of offheap regions. Plus I'd like to see results soon, your
proposal is
a major change for Ignite that can't be implemented fast enough.
Anyway, I think this makes sense, considering that one day U
Ivan, I think that we should use mmap/munmap to allocate huge chunks of
memory.
I've experimented with JNA and invoke mmap/munmap with it and it works fine.
May be we can create module (similar to direct-io) that use mmap/munap on
platforms, that support them
and fallback to Unsafe if not?
чт, 23
Hello Igniters,
I'd like to discuss the current issue with "out of memory" fails on
TeamCity. Particularly suites [1]
and [2], they have quite a lot of "Exit code 137" failures.
I investigated the "PDS (Indexing)" suite under [3]. There's another
similar issue as well: [4].
I came to the conclusi