Dear All,
after migrating from our 7mode FAS2040 cluster to CDot 8.3RC1 FAS2240 cluster, users are complaing about slow file access via NFS and CIFS. I ran sysstat on the node that hosts the affected vserver Corporate in the cluster and yes the CPU usage is quite high >70% most of the time.
GEDASAN::*> run GEDASAN-02 sysstat -x 1
CPU NFS CIFS HTTP Total Net kB/s Disk kB/s Tape kB/s Cache Cache CP CP Disk OTHER FCP iSCSI FCP kB/s iSCSI kB/s
in out read write read write age hit time ty util in out in out
67% 699 2 0 945 2289 60319 63500 24 0 0 1 88% 0% - 100% 210 0 34 0 0 22 5718
75% 964 0 0 1139 1512 46122 48228 8 0 0 1 88% 0% - 93% 93 0 82 0 0 117 5681
75% 1430 2 0 1476 2115 8835 13070 0 0 0 1 92% 0% - 88% 0 0 44 0 0 16 6111
67% 1225 0 0 1349 2208 69783 77632 24 0 0 1 91% 0% - 100% 60 0 64 0 0 303 7180
70% 942 0 0 1153 1992 56612 57096 0 0 0 1 91% 0% - 99% 146 0 65 0 0 99 7557
68% 1323 0 0 1378 3285 40221 52512 3712 0 0 1 93% 27% Ts 100% 0 0 55 0 0 71 5444
75% 1254 0 0 1298 1776 20286 36604 50804 0 0 1 93% 100% :f 100% 1 0 43 0 0 1 5882
77% 1459 0 0 1499 2298 13485 18304 6788 0 0 1 90% 49% : 100% 2 0 38 0 0 30 4895
68% 1231 1 0 1274 2008 45233 48920 0 0 0 1 91% 0% - 100% 0 0 42 0 0 1 5255
71% 1299 0 0 1414 2470 50187 49448 24 0 0 1 89% 0% - 100% 80 0 35 0 0 62 6267
69% 1171 1 0 1285 2418 40194 44296 0 0 0 1 87% 0% - 100% 80 0 33 0 0 22 5136
What looks strange to me is this:
*Total Read Write Other Read Write Latency
Vserver Ops Ops Ops Ops (Bps) (Bps) (us)
------- ------ ---- ----- ----- ------- ------- -------
VMWARE 1111 270 774 66 3966592 8445993 3672
Corporate
1090 7 49 1033 152174 763286 11825
VMWARE-PROJECT
449 87 290 71 1342976 1664991 5982
ISCSI 272 180 32 59 11021440
226560 11245
VMWARE-DMZ
231 33 159 38 818432 2506786 2734
VMWARE-VDI
69 9 40 19 5888 279218 2451
VMWARE-S1PDGS
20 0 18 1 0 60807 2465
ISCSI-DMZ
0 0 0 0 0 0 -
Other OPS for the affected vServer is really high. What is this? I could not find any information about it?
Best Regards,
Oliver
