I/O Analyzer 1.6.2 appliance doesn't generate reports.
Logs indicate:
unable to call charting scripts: Cannot run program "/bin/sh" (in directory "/root/ioAnalyzerScripts"): java.io.IOException: error=13, Permission denied.
ioAnalyzerScripts are in the /root directory which Tomcat cant get to.
As a workaround you can chmod 755 /root directory.
Hello,
I would like to know if there is a simple solution not to use the root account on the ESXi but a privileged account with a different name (I have an account called admin but no direct access to the root password itself.
Thanks
Attached are two VSAN performance testing profiles for the I/O Analyzer appliance.
They will be included in future releases of the I/O Analyzer.
They can be used with the VMware Virtual SAN POC Guide.
How to install them:
| (1) | Upload the ICF config files to ALL of I/O Analyzer controller AND worker VMs under their “/var/www/configs/” directory. |
| (2) | Restart the controller VM |
Exactly what does the "serialnumber" in the output from vscsiStats -e refer to?
Thanks
Hallo,
how can I restrict access even in the web interface ? Once you're in the console anyone can access .
Is there any way to request credentials from the web interface ?
Thanks and regards
Hi,
Does anyone know if the figures that make up the SUM are just maximums that have been hit or a constant throughout the whole test?
For instance in the io analyzer report the figures for ESX top show 900 read iops for two machines and a sum of 1800. Were the figures that make up that sum just taken from a point in time or the maximum hit?
Thanks
so here's my setup
I have deployed to the IO analyzer ova to 3 VMs
1 controller and 2 workers
on the controller I have created custom Iometer workloads, gave them proper names and placed them in
/var/www/configs/
then i did
/etc/init.d/tomcat6 restart
when I load up the analyzer webui, my custom workload shows up in the dropdown menu
next I proceeded to add my workers using my custom workload
but when I hit run and check each console session, IOmeter only loads up on the controller but not the 2 workers..
did I miss something?
thanks
Hi
I want to get the chain length of VMs and am struggling to find where in an API or even existing web tool this property is exposed.
I tried http://velemental.com/2012/05/05/unofficial-vmware-vcd-cmdlets/
Get-CIVMDiskFileChain but this did not work.
In the Get-CICM object there is an ExtensionData->Link but i can't find any documentation on this.
Context: VCD 5.51 and vCenter v5.51
Thanks in advance
Hello -
Our team is trying to use IOAnalyzer, however, it appears all of the I/O is hitting disk1 as opposed to the second disk that we've configured. Our setup is as follows:
1 Master with 8 workers.
Each appliance has disk 1 (sda) configured to reside on local storage of the host.
Each appliance has disk 2 (sdb) configured to reside on an iSCSI LUN.
After running the workload, the write/read requests are hitting the local disk (mpx.*) of the ESXi hosts. The iSCSI LUN (eui.*) show no I/O. Are we configuring the IO analyzers incorrectly? Do disk1 and disk2 need to reside on the same datastore?
Thanks in advance for the help.
Hello -
I recently setup an environment utilizing 8 IO Analyzers each accessing either an iSCSI LUN or NFS Share (but not both at the same time). The secondary virtual disk was set at 30GB.
For the iSCSI tests, we ran a 50/50 0% random workload and our total IOPS reached 4086.11.
When we created the second disk on an NFS datastore and ran the same test as above, our total IOPS only reached 583.57 for the same time period (2 hours). Additionally, latency doubled.
I checked IOMeter on the guests, and it appeared they were pushing fewer IOPS as well.
Any ideas as to why we couldn't push as many IOPS using NFS? I would think the amount of IOPS would be the same across tests, regardless of the backend?
Thanks in advance.
For I/O Analyzer worker VMs, it's recommended to create a larger second virtual hard disk which serves as the IOmeter target.
Is the size of the virtual Hard Disk exactly the same as the workload's working set size, that is, that all of the data in the virtual disk is read or written to storage during the run?
Do you have any other recommendations in deciding how large the second virtual hard disk should be for a worker VM?
Thank you,
Rebecca
Hello -
I have 10 IO Analyzer VMs setup (1 Master and 9 workers). As of late, we've been experiencing odd behavior with respect to the execution of workloads on the worker VMs. The scenario is as follows:
1. Schedule a workload for each worker VM
2. Master web page says the workload is being executed, however, the workers display no signs of IOMeter execution.
3. Reboot Workers - Upon startup, workers begin executing the workload that was previously scheduled.
Is there something I can do to reset the state of the workers without having to reboot? Is there a way to cancel tests in the master web page (or via CLI)?
Hi,
When I deploy this ova the vmware client complains, right at the end of the process, that the vmdk file fails the integrity check. Can someone kindly compare the md5sum for a working deployment of the ova? On my system it comes out as :
8a0e52bd575dab9f4c2830dac6e6c1ac vmware-io-analyzer-1.6.2.ova
I've tried chrome and firefox - both download the same content with the same md5sum. I've extracted the files from the zip via both Windows 7 Explorer and ark on opensuse 13.1.
Cheers,
Ron
Hi,
I finally got the v1.6.2 ova to deploy by using the WebClient instead of the NativeClient, however I went to boot it up and it will not boot. The error message is "No operating system found" and the vm's settings indicate that it is using the paravirtual scsi hba. I tried changing to the LSI Logic Parallel but that got the same result. Can anyone confirm what it should be? And has anyone been successful with v1.6.2?
Addendum : the deployment seems flawed. During the deployment wizard it states that two hard disks will be deployed - an 8GB Thin and a 16GB Thick/Lazy. However after deployment has completed there is a hard disk 1 - 16GB Thick and a hard disk 2 - 100MB Thick. I did choose Thick Eager for the disks during deployment but did not (could not) modify disk sizes. So far not having much luck with this fling...
Cheers,
Ron
Hello,
I am trying out IO Analyzer for the first time and I believe I have everything set up correctly but in the Web UI I cannot get a workload added.
I when I deployed the workers I deleted the second HD and added my new target and selected them as Thick Provisioned Eager Zeroed. I added my hosts to the UI and I can see my individual worker VMs but when I select the VM, Test, and Workload the Add button does nothing. From what I have seen I should see this workload move down to the workload configuration pane correct? Any ideas on what I might be doing wrong?
Thanks
Eric
Hi, just wanted to clarify something regarding the trace files. I am right to assume that the last field, latency, is ignore during playback or does it have a part to play?
Thanks
Vscsistats -e trace output is like this,
#Command Type,Serial Number,IO Data Length,Num SG Entries,LBN,Time Stamp (microseconds),Command Latency (microseconds)
write,2148401171,4096,1,42447504,10890477934180,69309
write,2148401179,4096,1,42722448,10890477979970,40508
write,2148401171,8192,1,20015856,10890478013739,21616
write,2148401171,401408,1,20015872,10890478142128,9076
Timestamp (microseconds) is epoch in above case [if that's true then it does not match with current date], can anyone provide description or link which describes "Vscsi Command trace" of vscsistats (-t option) command?
I deployed "vmware-io-analyzer-1.6.2.ova" on two ESXi hosts say as ESX1 and ESX2. followed all instructions mentioned in user guide
I did not see any workers added on iometer.
I am not understanding how it is working.
Please find attachments for logs.
It will be good if i get some help on it.
Thanks
~Kalyani
Hello,
I believe it would make sense to open the sources to the community. GO opensource please.
Please get rid of the IOMeter Tool. Yes IOmeter is a great tool but the alternative fio provides more.
Use the community to evolve it. What about vCloud Air integration, or make use of plugable provider to stage VMs to other cloud solutions or plattforms. Or to stage docker ontop of vsphere. ... etc.
cheers,
Thomas
I have freshly downloaded, and deployed the OVA. It's working just fine on the controller, but we do not get the guests to run their workloads. I've worked through the log files, and scripts, and think the problem occurs with an empty script
localhost:~/ioAnalyzerScripts # ls -al | grep resxtop
-rwxr-xr-x 1 root root 0 Dec 16 2013 run-resxtop.sh
I have tried to re-deploy, and look at each of the VMs for this file, and they're all empty.
I traced this down to line 65 in run-experiments-all.pl. The configuration files before get copied across via scp without issue, and the code after executes trying to pick up the results and then runs through 10 timeouts.
Is there a way to get a working copy of this file? I downloaded the tarball as well, but am stuck with only a single worker.