[zfs-discuss] Please explain this "zpool iostat" write bandwidth on a read-only pool

Cédric Lemarchand cedric.lemarchand at ixblue.com
Tue Dec 16 15:19:25 EST 2014


The read only flag seems to only concern data, not metadata, I find it s a plausible explanation. 

--
Cédric Lemarchand

> Le 16 déc. 2014 à 21:16, Michael Kjörling <michael at kjorling.se> a écrit :
> 
> On 16 Dec 2014 14:02 -0600, from lkateley at kateley.com (Linda Kateley):
>>> All this to say: **what's that 37 write bandwidth on "anatolian"**
>>> which is supposedly read-only? It seems to be holding reasonably
>>> constant.
>> 
>> Try turning off atime. It is probably just writing updates to atime.
> 
> That's odd in more ways than one. atime was turned on so I set it to
> off on the anatolian root file system (no reason for atime=on on a
> backup pool, as I see it), which succeeded. That, however, if
> anything, appears to have _increased_ the write bandwidth: it now
> seems to be holding at _41_ rather than 37. It still doesn't _explain_
> the behavior: why would ZFS do atime writes to an explicitly
> _read-only_ file system? And why does it not show as write
> "operations"?
> 
> -- 
> Michael Kjörling • https://michael.kjorling.semichael at kjorling.se
> OpenPGP B501AC6429EF4514 https://michael.kjorling.se/public-keys/pgp
>                 “People who think they know everything really annoy
>                 those of us who know we don’t.” (Bjarne Stroustrup)
> 
> To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+unsubscribe at zfsonlinux.org.

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+unsubscribe at zfsonlinux.org.



More information about the zfs-discuss mailing list