VY
2018-02-20 22:08:45 UTC
Dear All
We have a Linux host with a MYSQL instance that runs on an NFS mounted disk
drive. I want to understand/tune the NFS mount such that I can get it to as
performant as possible. Below is the output of "nfsiostat".
My questions
1). "retrans" for both read and write is a not a small number.
Is that good or bad? I would expect it is best to minimize that number but
like to hear opinion
2). "ave exe" is between 2 to 5ms. Isn't that a bit high? I have MySQL
running on it and if there's delay with the filesystem that will have some
impact to MySQL?
Any tips on how to read this data will be much appreciated!
-- nfsiostat --
op/s rpc bklog 2825.86 0.00
read: ops/s kB/s kB/op retrans avg RTT (ms) avg exe (ms) 1156.465 23124.440
19.996 18845 (0.0%) 0.962 2.328
write: ops/s kB/s kB/op retrans avg RTT (ms) avg exe (ms) 1223.878
27121.697 22.160 364917 (0.0%) 0.623 4.222
We have a Linux host with a MYSQL instance that runs on an NFS mounted disk
drive. I want to understand/tune the NFS mount such that I can get it to as
performant as possible. Below is the output of "nfsiostat".
My questions
1). "retrans" for both read and write is a not a small number.
Is that good or bad? I would expect it is best to minimize that number but
like to hear opinion
2). "ave exe" is between 2 to 5ms. Isn't that a bit high? I have MySQL
running on it and if there's delay with the filesystem that will have some
impact to MySQL?
Any tips on how to read this data will be much appreciated!
-- nfsiostat --
op/s rpc bklog 2825.86 0.00
read: ops/s kB/s kB/op retrans avg RTT (ms) avg exe (ms) 1156.465 23124.440
19.996 18845 (0.0%) 0.962 2.328
write: ops/s kB/s kB/op retrans avg RTT (ms) avg exe (ms) 1223.878
27121.697 22.160 364917 (0.0%) 0.623 4.222