Description of problem: iostat -n shows exactly the same values for rops and wops Version-Release number of selected component (if applicable): 7.0.2-3 How reproducible: iostat -kxn 2 (with NFS volumes mounted) Steps to Reproduce: 1. mount an NFS volume 2. cp a file from this volume to local storage 3. iostat -kxn 2 Actual results: Device: rkB_nor/s wkB_nor/s rkB_dir/s wkB_dir/s rkB_svr/s wkB_svr/s rops/s wops/s F3040:/vol/vol0 0.00 0.00 0.00 0.00 0.00 0.00 5627.00 5627.00 F3040:/vol/ed 20018.00 0.00 0.00 0.00 19920.00 0.00 5627.00 5627.00 Expected results: Device: rkB_nor/s wkB_nor/s rkB_dir/s wkB_dir/s rkB_svr/s wkB_svr/s op/s rops/s wops/s F3040:/vol/vol0 0.00 0.00 0.00 0.00 0.00 0.00 7594.00 0.00 0.00 F3040:/vol/ed 26986.00 0.00 0.00 0.00 26928.00 0.00 7594.00 841.50 0.00 Additional info: Although my expected results came from a prototype version I had (where op/s was dropped in the production version), it seems like ops/s is being used instead of rops and wops; only a guess, though. F3040 is a Netapp 3040 mount arguments: rw,hard,nointrf,vers=3 rsize=32768,wsize=32768,proto=udp
Please could you attach here the output of: cat /proc/self/mountstats | grep xprt for me it seems iostat is ok - but I need this data.
So, I retested it with a simple copy from the mount to a local mount: old version (a prototype I have; never released with 5.3) Device: rkB_nor/s wkB_nor/s rkB_dir/s wkB_dir/s rkB_svr/s wkB_svr/s op/s rops/s wops/s F960:/vol/vol0/home 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 F960:/vol/amd 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 F3040:/vol/vol0 0.00 0.00 0.00 0.00 0.00 0.00 8005.50 0.00 0.00 F3040:/vol/ed 28482.00 0.00 0.00 0.00 28320.00 0.00 8005.50 885.00 0.00 RHEL5.3 production version Device: rkB_nor/s wkB_nor/s rkB_dir/s wkB_dir/s rkB_svr/s wkB_svr/s rops/s wops/s F960:/vol/vol0/home 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 F960:/vol/amd 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 F3040:/vol/vol0 0.00 0.00 0.00 0.00 0.00 0.00 9186.50 9186.50 F3040:/vol/ed 36682.00 0.00 0.00 0.00 482.00 0.00 9186.50 9186.50 xprt: tcp 1013 0 8 3 69 39 39 0 39 0 xprt: tcp 1013 0 8 3 69 39 39 0 39 0 xprt: tcp 891 0 8 3 43 192645 192645 0 325447 0 xprt: tcp 891 0 8 3 43 192645 192645 0 325447 0 During a copy only (i.e. no writes to the Netapp), there should be 0 wops and in this case, they are identical (implying that rops and wops are reading the same value and the wrong value). My guess is that rops and wops are being mapped to ops. If you had collapsed rops and ops together, then 91xx would be the right value. However, wops should remain at zero. I would really like this fixed for 5.3 as some customers would like to use this feature (Yahoo, Shopzilla, Edmunds).
Hello, yes in this case there is used wrong value - this problem is fixed in the upstream version , here could be used the same fix.
This request was evaluated by Red Hat Product Management for inclusion, but this component is not scheduled to be updated in the current Red Hat Enterprise Linux release. If you would like this request to be reviewed for the next minor release, ask your support representative to set the next rhel-x.y flag to "?".
I really need this fixed now in version 5.3. What is needed to make that happen?
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: Previously, the "iostat -n" command used invalid data to create statistics for read and write operations. With this update, the data source for these statistics has been fixed, and the iostat utility now returns correct information.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2011-1005.html