On 01/22/2014 03:06 PM, babajaga wrote:
IOs have a variable size and for writing an object to a file with the aufs
store,
the OS write meta data to the file system log, updates the inode table and
writes the data to a new file.
So for aufs for one logical 'write object to disk' there are 3 IOs.
I do not know the internals of rock fs but most likely only does one IO for
each 'write object to disk'. <
I do not think, that this is correct. For aufs it might be 3 "logical" IOs,
which mean,s from the standpoint of view of application (i.e.squid/aufs).
But because LINUX does a lot of bufffering/merging, in very rare scenarios
these 3 "logical" IOs will be real "physical"IOs (disk accesses).
The 3 mentioned I/Os are for log, inode and data. Multiple logical Log IOs can be combined into one physical IOs,
logical inode IOs can be combined into physical IOs, data IOs of different files are rarely combined.
I do not want to discuss file system IOs into too much detail here, the squid group is not meant for that.
The important thing is about understanding that aufs and rock fs are different and
cause different IO patterns.
For rock, the one "logical"IO does not necessarily result in one "physical"
IO either.
So, depending upon the size of the necessary storage, the amount of RAM, and
tweaking the LINUX-fs, the number of physical IOs is much lower compared to
the number of "logical" IOs, squid/aufs executes.
(ext4 for example has a lot of knobs for performance tuning)
--
View this message in context: http://squid-web-proxy-cache.1019090.n4.nabble.com/Hypothetically-comparing-SATA-SAS-to-NAS-SAN-for-squid-tp4664350p4664381.html
Sent from the Squid - Users mailing list archive at Nabble.com.
