7:03 a.m.
Hi,
I have a requirement in my FSM to invalidate the cache.
While doing it I want to nfs-ganesha to drop any in-memory cache for the given export.
How can I inform nfs-ganesha to drop all in-memory caches for the given export?
Regards,
Shahid Shaikh.
9:07 a.m.
New subject: FSAL cache cleanup - How to drop in-mem caches at nfs-ganesha
Currently, there's no way to do this. You can invalidate the metadata
cache for a single object with an UP-call (the invalidate() call) but
there is no way to recursively invalidate, and no UP-call to remove a
handle-cache entry at all. This last would chancy anyway, since
anything with a ref on that entry would keep it around regardless of any
UP-call.
However, once the underlying inode for an entry is gone, the first use
of it will return ESTALE, causing the entry to be removed ASAP.
Daniel
On 07/25/2018 10:03 AM, sshaikh(a)commvault.com wrote:
Hi,
I have a requirement in my FSM to invalidate the cache.
While doing it I want to nfs-ganesha to drop any in-memory cache for the given export.
How can I inform nfs-ganesha to drop all in-memory caches for the given export?
Regards,
Shahid Shaikh.
_______________________________________________
Devel mailing list -- devel(a)lists.nfs-ganesha.org
To unsubscribe send an email to devel-leave(a)lists.nfs-ganesha.org
10:44 p.m.
New subject: FSAL cache cleanup - How to drop in-mem caches at nfs-ganesha
Thanks Daniel.
How long nfs-ganesha will wait before validating the cache for given gid against the
underlying inode?
6:59 a.m.
New subject: FSAL cache cleanup - How to drop in-mem caches at nfs-ganesha
The next time that inode is used. There's no background validation.
Daniel
On 07/26/2018 01:44 AM, sshaikh(a)commvault.com wrote:
Thanks Daniel.
How long nfs-ganesha will wait before validating the cache for given gid against the
underlying inode?
_______________________________________________
Devel mailing list -- devel(a)lists.nfs-ganesha.org
To unsubscribe send an email to devel-leave(a)lists.nfs-ganesha.org
2093
days inactive
2095
days old
3 comments
2 participants
participants (2)
-
Daniel Gryniewicz -
sshaikh＠commvault.com