Question: the VSZ of pantsd shows up as enormous v...
# general
f
Question: the VSZ of pantsd shows up as enormous values: like 539 G in my case
Copy code
3642272 jreed      20   0  539G 92532 25036 S  0.0  0.3  0:00.00 pantsd [/home/jreed/.../repo-root]
Is this related to mmap or something used in caching? Or is something mapped in memory being counted multiple times
checked
/proc/$id/maps
and it does indeed seems to be tied to huge maps allocated by lmdb_store ... I guess this is just an artifact of all that
w
yep.
LMDB is very simple, which is nice. but we’ve had to shard it, which multiplies the VM usage by 16 by default… and loses some of the simplicity benefit.
🙏🏻 1