Performance Monitor and % Disk Time

-

measuring % disk time in performance monitor of virtual machine known inaccurate?

i'm testing virtualisation of citrix servers. the default resource manager metrics in citrix examine %disk time. throwing warnings hitting values of 80,000% , greater (averaging around 600%). these same counters performance monitor. can switch off in citrix avoid warnings i'm curious should observed (as don't same extremes under vmware). 

the %disk time observed on host o/s negligible doesn't appear there real problem disk performance. scale of perfmon disk counters out factor of 1000 (or more).

i've set several different configurations example:
hp dl360 g5 2x 72gb mirrored root partition and 2x146gb mirrored vhd's 
host o/s windows server 2008 x64 latest patches (including hyper-v update)
guest o/s windows server 2003 x86 running citrix ps 4.0 - configured fixed disk

hello,

 

based on understanding, want use performance monitor inside vm, not on parent partition, right?

 

well, these counters may yield unlikely counter values. here summary quoted following article:

 

"when working disk-time or disk-queue length counters, aware of following limitations might yield unlikely counter values.

 

the % disk read time , % disk write time counters can exaggerate disk time. because report busy time based on duration of i/o request, includes time spent in activities other reading or writing disk. sums busy time requests , divides elapsed time of sample interval. if multiple requests in process @ time, total request time greater time of sample interval; result, reported disk utilization can exceed actual utilization.

 

counter values report sums can misleading multidisk systems. when @ _total instance % disk time or disk-queue counters on multidisk system, counters report values totaled disks , not divide these totals on number of disks in use. therefore, in system 1 idle disk , 1 disk 100 percent busy, can appear if disks 100 percent busy."

 

this may understand why value exceeding 100%. suggest use following counters instead of % disk counters:

 

avg. disk read queue length

avg. disk bytes/read

avg. disk sec/read

disk read bytes/sec

disk reads/sec

 

all these information same windows server 2003 , windows server 2008, documented in following articles (for windows 2000):

 

working disk counters:

http://www.microsoft.com/technet/prodtechnol/windows2000serv/reskit/prork/pree_exa_swlz.mspx?mfr=true

 

monitoring disk efficiency:

http://www.microsoft.com/technet/prodtechnol/windows2000serv/reskit/prork/pree_exa_qkkb.mspx?mfr=true

 

i hope can of help.

 

best regards,

chang yin


Windows Server  >  Hyper-V



Comments

Post a Comment

Popular posts from this blog

some help on Event 540

-
does event 540 me distinguish between user , administrator access?    this event generated ypon successfull logon, logon exactly? successful logon on system?   is there other way log administrator access system using event viewer?   thank valuable response   georges hi,   as far know, cannot distinguish between user , administrator access event 540. more information, please refer to:   event id: 540 http://www.microsoft.com/technet/support/ee/transform.aspx?prodname=windows+operating+system&prodver=5.0&evtid=540&evtsrc=security&lcid=1033   events 528 , 540 http://blogs.msdn.com/b/ericfitz/archive/2004/12/09/279282.aspx   regards, bruce this posting provided "as is" no warranties, , confers no rights. please remember click "mark answer" on post helps you, , click "unmark answer" if marked post not answer question. can beneficial other community members reading thread.
Read more

WMI Repository 4GB limit - Win 2003 Ent Question

-
Image
hi there, i've question size limit on wmi repository (being 4gb). if isn't correct place post this, apologies , can let me know wheere post this? i've come across few servers running citrix have reached limit on wmi repository size , need install ibm director agent hw monitoring. during install getting 'out of disk space' errors entered in mofcomp.log: "(mon apr 02 04:10:36 2012.172843) : error occurred while processing item 4: (mon apr 02 04:10:36 2012.172843) : error number: 0x8004103b, facility: wmi description: out of disk space" when error number http://msdn.microsoft.com/en-us/library/windows/desktop/aa394559%28v=vs.85%29.aspx i see this: wbem_e_out_of_disk_space 2147749947 (0x8004103b) disk out of space or 4 gb limit on wmi repository (cim repository) size reached. windows xp , windows 2000/nt:   disk out of space. i'm wondering if there reason why limit set 4gb or can increase needs be? can't seem find reas
Read more

Event ID 1302 (error 1307) DFS replication service encountered an error while writing to the debug log file

-
Image
hello. we @ step 0 of migration frs dfsr sysvol replication on windows 2008r2 dc. every time run powershell dfsrmig / getmigrationstate says: "unable create dfsr migration log file. error 1307. all domain controllers have migrated global state 'start'. migration has reached consistent state on domain controllers. succeeded." it creates event id 1302 in "dfs replication" event log, explaining error 1307 means: "this security id may not assigned ownder of object ". i already checked space limits, quotas, permissions on c:\windows\debug according microsoft support article unable fix it. although, there new entries added dfsr00095.txt log file despite error. dfsrxxxx file has not been archived .gz format, and, beleive, contains dfsr diagnostics , events. here of latest entries: 20150908 08:47:27.867 11616 sysm  3354 migration::sysvolmigration::getsysvolreadyflag [mig] sysvol ready 20150908 08:47:27.867 11616 sysm   456 migrati
Read more