Stats should show counters for most rows. See example below: # cat /proc/fs/cifs/Stats Resources in use CIFS Session: 1 Share (unique mount targets): 2 SMB Request/Response Buffer: 1 Pool size: 5 SMB Small Req/Resp Buffer: 1 Pool size: 30 Operations (MIDs): 0 0 session 0 share reconnects Total vfs operations: 36 maximum at one time: 2 Max requests in flight: 3 1) \\localhost\test SMBs: 67 Bytes read: 90177536 Bytes written: 2 Open files: 0 total (local), 0 open on server TreeConnects: 1 total 0 failed TreeDisconnects: 0 total 0 failed Creates: 12 total 0 failed Closes: 13 total 1 failed Flushes: 1 total 0 failed Reads: 22 total 0 failed Writes: 1 total 0 failed Locks: 0 total 0 failed IOCTLs: 1 total 1 failed QueryDirectories: 2 total 0 failed ChangeNotifies: 0 total 0 failed QueryInfos: 12 total 0 failed SetInfos: 2 total 0 failed OplockBreaks: 0 sent 0 failed Are you running an old kernel (pre-5.0 e.g.)? On Thu, May 28, 2020 at 3:39 PM Matthias Leopold <matthias.leopold@xxxxxxxxxxxxxxxx> wrote: > > Hi, > > I'm trying to debug the performance of rsync reading from a Windows 2012 > R2 share mounted readonly in CentOS 7. I tried to use cifsiostat, which > doesn't print any stats. I looked into /proc/fs/cifs/Stats and saw that > it contains mostly "0" for counters (I would expect to see some numbers > for eg "Reads"). What am I doing wrong? > > options from /proc/mounts are > ro,relatime,vers=3.0,cache=strict,username=foo,domain=xxx,uid=1706,forceuid,gid=1676,forcegid,addr=10.110.81.122,file_mode=0660,dir_mode=0770,soft,nounix,serverino,mapposix,rsize=61440,wsize=1048576,echo_interval=60,actimeo=1 > > thx > Matthias -- Thanks, Steve
