T(write) = T(position) + (D/R(peak))
R(effective) = D / T(write) = D / T(position) + (D/R(peak)) = F * R(peak)
D = R(effective) * T(write)
D = F / (1-F) * R(peak) * T(position)
T(write) = T(position) + (D/R(peak))
R(effective) = D / T(write) = D / T(position) + (D/R(peak)) = F * R(peak)
D = R(effective) * T(write)
D = F / (1-F) * R(peak) * T(position)
80900
(one two three four five six seven)
cleaning the log - LFS
we expect the cleaner to read in M existing segments, compact their contents into N new segments (where N < M), and then write the N segments to disk in new locations. The old M segments are then freed and can be used by the file system for subsequent writes.
segment summary block : to determine whether a block is live or dead.
READ and LOOKUP - partially idempotent
WRITE - idempotent
GFS:
The master periodically communicates with each chunkserver in HeartBeat messages to give it instructions and collect its state
Clients interact with the master for metadata operations, but all data-bearing communication goes directly to the chunkservers
a client asks the master which chunkservers it should contact
Virual File System:
The key idea is toabstract out that part of the file system that is common to all file systems and putthat code in a separate layer that calls the underlying concrete file systems to actually manage the data.
Congrats
New AT 🥳
So kafi achi parhai ki Hai masu ne idhar😃
Good going😃
yes im gonna continue studying here only.. sorry replies baad me
-no caching of data at client end
-master doesn't keep log of chunkservers having replicas
-order of append doesn't matter
-stale replicas -> replica chunks and master keeps updating version number -> if a chunk doesn't have updated number -> master detects it as stale 😲
-garbage collection -> hides files instead of deleting -> deletes them only if they arent reaccessed for 3 days
-
comment:
p_commentcount