2:55 p.m.
Sagar Singh has uploaded this change for review. (
https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/540965 )
Change subject: Trying to reap the MDCache LRU entries traversing through the whole queue,
instead of checking only the first LRU entry of every lane.
......................................................................
Trying to reap the MDCache LRU entries traversing through the whole queue, instead of
checking only the first LRU entry of every lane.
When lru entries used are greater than high watermark then lru_reap_impl() try
to reap the first entry of LRU queue for given qid, going through every lane.
If the refcount is not (LRU_SENTINEL_REFCOUNT + 1), then we are leaving the whole lane, in
which
There are many entries which can be reaped. So this is slowing down the reaping and
causing high
memory use by the ganesha server. This scenario was resulting to run OOM Killer, so
ganesha server
was crashing with a segmentation fault.
This patch is written in the hope that fast LRU reaping will not cause high memory usage
by the MDCache.
Change-Id: I2a2fae830ce678b8004ac80aa648b6cbffab77cf
Signed-off-by: Sagar Singh <sagar.it.iiests(a)gmail.com>
---
M src/FSAL/Stackable_FSALs/FSAL_MDCACHE/mdcache_lru.c
M src/FSAL/Stackable_FSALs/FSAL_MDCACHE/mdcache_lru.h
2 files changed, 135 insertions(+), 1 deletion(-)
git pull ssh://review.gerrithub.io:29418/ffilz/nfs-ganesha refs/changes/65/540965/1
--
To view, visit https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/540965
To unsubscribe, or for help writing mail filters, visit
https://review.gerrithub.io/settings
Gerrit-Project: ffilz/nfs-ganesha
Gerrit-Branch: next
Gerrit-Change-Id: I2a2fae830ce678b8004ac80aa648b6cbffab77cf
Gerrit-Change-Number: 540965
Gerrit-PatchSet: 1
Gerrit-Owner: Sagar Singh <sagar.it.iiests(a)gmail.com>
Gerrit-MessageType: newchange