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@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@gmail.com> Gerrit-MessageType: newchange