538 reputation
922
bio website yz.mit.edu
location
age
visits member for 4 years, 8 months
seen Mar 22 at 4:09

Nov
9
awarded  Notable Question
Oct
27
awarded  Self-Learner
Sep
26
awarded  Notable Question
Aug
27
awarded  Notable Question
Aug
24
answered nginx proxied responses terminating/truncating
Aug
24
revised nginx proxied responses terminating/truncating
added 375 characters in body
Aug
24
asked nginx proxied responses terminating/truncating
Aug
11
awarded  Popular Question
Aug
11
comment ZFS on >2TB disk in Ubuntu 12.04?
@MichaelKjörling Well, unless I got scammed by Fry's, yeah it's a WD 4TB drive. I heard the 2TB capping is what HPA does. See askubuntu.com/questions/330666/zfs-on-2tb-disk-in-ubuntu-12-04/…
Aug
9
revised ZFS on >2TB disk in Ubuntu 12.04?
deleted 3 characters in body
Aug
9
asked ZFS on >2TB disk in Ubuntu 12.04?
Jun
23
accepted SSL connection errors from Apache
Jun
12
comment OOM despite available memory (cache)
@CraigRinger Unfortunately there are other considerations including sticking with the kernel in the Ubuntu 12.04 distro for LTS, compatibility, security updates, etc. We're just interested if someone knows how to explain what's happening here - if it helps, pretend we're not interested in changing the status quo / making the problem go away. BTW shared_buffers is still in PG9.3.
Jun
12
comment OOM despite available memory (cache)
After reviewing the links, there are unfortunately a number of assertions without backing evidence or concrete technical explanations. There certainly many Linux environments where swap is not even an option (example: look no further than a Live CD where there isn't already an existing local swap partition to reuse). We are furthermore not interested in enabling swappiness based on our own experiences and environment - we would rather have the OOM. An answer to the original question would be appreciated.
Jun
12
comment OOM despite available memory (cache)
Thanks for the elaboration on swap, but this doesn't answer my question of why this is happening in the first place. Yes, I understand the basic premise that Linux overcommits by default and that OOM is when we're out of RAM - I could have stated this much in my original question. But the question is why is it kicking in when I still have plenty of RAM (much of it just sitting in the FS cache)? Assume I'm not even interested in changing anything - that the OOM killer is fine, so long as I understand why it's being triggered.
Jun
12
comment OOM despite available memory (cache)
Related chat discussion
Jun
11
asked OOM despite available memory (cache)
Jun
7
comment Details on ZFS scrubbing repairs requirements/guarantees
"relocating blocks elsewhere" - does ZFS thereafter permanently blacklist the original location of the blocks?
Jun
7
accepted Details on ZFS scrubbing repairs requirements/guarantees
Jun
6
asked Details on ZFS scrubbing repairs requirements/guarantees