I was wondering if there is documentation that details the meanings of
everything in /proc/net/rpc/nfsd
I know what th is, but I am curious what other information I can glean
from the file.
Thanks for your help,
--
Matt Schillinger
[email protected]
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
The only event dedicated to issues related to Linux enterprise solutions
http://www.enterpriselinuxforum.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
Statistics for the request reply cache:
rc <hits> <misses> <nocache>
Statistics for filehandle lookup:
fh <stale> <total-lookups> <anon-lookups> <dir-not-in-dcache>
<nondir-not-in-dcache>
Statistics for IO throughput:
io <bytes-read> <bytes-written>
time (seconds) when nfsd thread usage above thresholds and number of
times that all threads were in use:
th <threads> <fullcnt> <10%-20%> <20%-30%> ... <90%-100%>
Number of times that read-ahead entry was found that deep in the cache:
ra <cache-size> <10% <20% <30% ... <100% <not-found>
Socket statistics:
net <netcnt> <netudpcnt> <nettcpcnt> <nettcpconn>
where:
<netcnt>
Number of received requests
<netudpcnt>
Number of received datagrams on UDP sockets
<nettcpcnt>
Number of received packets on TCP sockets
<nettcpconn>
Number of accepted TCP connections
RPC statistics:
rpc <rpccnt> <total-bad> <rpcbadfmt> <rpcbadauth> <rpcbadclnt>
where:
<rpccnt>
Number of RPC requests
<total-bad>
Total number of bad requests
<rpcbadfmt>
Bad format
<rpcbadauth>
Authentication failed
<rpcbadclnt>
N/A
NFSv2 server procedure statistics:
proc2 <nproc> <null> <getattr> <setattr> <root> <lookup> <readlink>
<read> <wrcache> <write> <create>
<remove> <rename> <link> <symlink> <mkdir> <rmdir> <readdir> <fsstat>
NFSv3 server procedure statistics:
proc3 <nproc> <null> <getattr> <setattr> <lookup> <access> <readlink>
<read> <write> <create> <mkdir>
<symlink> <mknod> <remove> <rmdir> <rename> <link> <readdir>
<readdirplus> <fsstat> <fsinfo> <pathconf>
<commit>
Regards,
---
Bruce Allan <[email protected]>
Software Engineer, Linux Technology Center
IBM Corporation, Beaverton OR
503-578-4187 IBM Tie-line 775-4187
Matt Schillinger
<[email protected] To: [email protected]
m> cc:
Sent by: Subject: [NFS] more info please
[email protected]
ceforge.net
05/14/2003 10:24 AM
I was wondering if there is documentation that details the meanings of
everything in /proc/net/rpc/nfsd
I know what th is, but I am curious what other information I can glean
from the file.
Thanks for your help,
--
Matt Schillinger
[email protected]
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
The only event dedicated to issues related to Linux enterprise solutions
http://www.enterpriselinuxforum.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
The only event dedicated to issues related to Linux enterprise solutions
http://www.enterpriselinuxforum.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
Bruce ...
How about NFSv4 statistics ??? ...
Duc
[email protected], Phone: (512) 838-9919 Fax: (512) 838-0070
Bruce
Allan/Beaverton/IBM@ To: Matt Schillinger <[email protected]>
IBMUS cc: [email protected]
Sent by: Subject: Re: [NFS] more info please
[email protected]
ceforge.net
05/14/2003 12:32 PM
Statistics for the request reply cache:
rc <hits> <misses> <nocache>
Statistics for filehandle lookup:
fh <stale> <total-lookups> <anon-lookups> <dir-not-in-dcache>
<nondir-not-in-dcache>
Statistics for IO throughput:
io <bytes-read> <bytes-written>
time (seconds) when nfsd thread usage above thresholds and number of
times that all threads were in use:
th <threads> <fullcnt> <10%-20%> <20%-30%> ... <90%-100%>
Number of times that read-ahead entry was found that deep in the cache:
ra <cache-size> <10% <20% <30% ... <100% <not-found>
Socket statistics:
net <netcnt> <netudpcnt> <nettcpcnt> <nettcpconn>
where:
<netcnt>
Number of received requests
<netudpcnt>
Number of received datagrams on UDP sockets
<nettcpcnt>
Number of received packets on TCP sockets
<nettcpconn>
Number of accepted TCP connections
RPC statistics:
rpc <rpccnt> <total-bad> <rpcbadfmt> <rpcbadauth> <rpcbadclnt>
where:
<rpccnt>
Number of RPC requests
<total-bad>
Total number of bad requests
<rpcbadfmt>
Bad format
<rpcbadauth>
Authentication failed
<rpcbadclnt>
N/A
NFSv2 server procedure statistics:
proc2 <nproc> <null> <getattr> <setattr> <root> <lookup> <readlink>
<read> <wrcache> <write> <create>
<remove> <rename> <link> <symlink> <mkdir> <rmdir> <readdir> <fsstat>
NFSv3 server procedure statistics:
proc3 <nproc> <null> <getattr> <setattr> <lookup> <access> <readlink>
<read> <write> <create> <mkdir>
<symlink> <mknod> <remove> <rmdir> <rename> <link> <readdir>
<readdirplus> <fsstat> <fsinfo> <pathconf>
<commit>
Regards,
---
Bruce Allan <[email protected]>
Software Engineer, Linux Technology Center
IBM Corporation, Beaverton OR
503-578-4187 IBM Tie-line 775-4187
Matt Schillinger
<[email protected] To:
[email protected]
m> cc:
Sent by: Subject: [NFS] more info
please
[email protected]
ceforge.net
05/14/2003 10:24 AM
I was wondering if there is documentation that details the meanings of
everything in /proc/net/rpc/nfsd
I know what th is, but I am curious what other information I can glean
from the file.
Thanks for your help,
--
Matt Schillinger
[email protected]
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
The only event dedicated to issues related to Linux enterprise solutions
http://www.enterpriselinuxforum.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
The only event dedicated to issues related to Linux enterprise solutions
http://www.enterpriselinuxforum.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
The only event dedicated to issues related to Linux enterprise solutions
http://www.enterpriselinuxforum.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
>>>>> " " == Duc Vianney <[email protected]> writes:
> How about NFSv4 statistics ??? ...
NFSv4 is a bit more difficult as there are only 2 RPC calls that are
defined: NULL and COMPOUND.
We're working around this to provide more detailed statistics, but for
the moment we're limited to the following:
<NULL> <COMPOUND> <READ> <WRITE> <COMMIT> <OPEN> <OPEN_CONFIRM>
<CLOSE> <SETATTR>
(where <COMPOUND> in the above list means: "all other operations")
Cheers,
Trond
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
The only event dedicated to issues related to Linux enterprise solutions
http://www.enterpriselinuxforum.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
Hi Trond,
Have the newer statistics counters been put in since 2.5.66? On 2.5.66 all
I see are 'proc4 <nproc> <#null> <#compound>'. It's good to know more
detail is being added; I really need to grab the latest bits.
Regards,
---
Bruce Allan <[email protected]>
Software Engineer, Linux Technology Center
IBM Corporation, Beaverton OR
503-578-4187 IBM Tie-line 775-4187
Trond Myklebust
<trond.myklebust@ To: Duc Vianney/Austin/IBM@IBMUS
fys.uio.no> cc: Bruce Allan/Beaverton/IBM@IBMUS, Matt Schillinger <[email protected]>,
[email protected], [email protected]
05/14/2003 02:56 Subject: Re: [NFS] more info please
PM
>>>>> " " == Duc Vianney <[email protected]> writes:
> How about NFSv4 statistics ??? ...
NFSv4 is a bit more difficult as there are only 2 RPC calls that are
defined: NULL and COMPOUND.
We're working around this to provide more detailed statistics, but for
the moment we're limited to the following:
<NULL> <COMPOUND> <READ> <WRITE> <COMMIT> <OPEN> <OPEN_CONFIRM>
<CLOSE> <SETATTR>
(where <COMPOUND> in the above list means: "all other operations")
Cheers,
Trond
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
The only event dedicated to issues related to Linux enterprise solutions
http://www.enterpriselinuxforum.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
>>>>> " " == Bruce Allan <[email protected]> writes:
> Hi Trond,
> Have the newer statistics counters been put in since 2.5.66?
> On 2.5.66 all I see are 'proc4 <nproc> <#null> <#compound>'.
> It's good to know more detail is being added; I really need to
> grab the latest bits.
Grab the latest bk snapshot. The NFS client on 2.5.x wasn't really
very stable prior to Tuesday's snapshot.
But yes: as you can see we are in the process of adding stuff to the
nfs4_procedures[] array. Those 'RPC procedures' will again provide the
more detailed statistics (and hopefully also faster+better NFSv4
performance).
Cheers,
Trond
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
The only event dedicated to issues related to Linux enterprise solutions
http://www.enterpriselinuxforum.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs