HPReg,
Thank you for the solution. I've put together a little program based on your code that does exactly what you described. That program is invoked by launchd at login and works like a charm. It's definitely a viable workaround for the issue and I appreciate your willingness to share it with me.
A thousand thanks.
alfine ce la feci.... ebbene sì: se usate un utente windows per il servizio vCenter LA SUA PASSWORD NON DEVE CONTENERE CARATTERI STRANI. Anche il "-" mi ha dato problemi.
Stessa regola vale per la password dell'amministratore dell'SSO, ma qui almeno è documentato.
Inoltre con la 5.5b viene aggiunta l'Identity Source Active Directory (se lo desiderate E se il server è in un dominio).
Però a differenza di quanto accadeva con la 5.x, ora è necessario loggarsi nel webclient una prima volta come administrator@vsphere.local e assegnare i permessi di amministratore del vCenter agli utenti desiderati.
Io avevo messo gli utenti nel gruppo Administrators della macchina e messo questo come Administrator del vCenter MA NON funzionava: ho dovuto aggiungerli esplicitamente uno per uno come Administrator del vCenter.
Attenzione anche alla questione della versione dell'SQL Native Client: la 5.5a non sembrava "schizzinosa" se la versione non era identica al server db, la 5.5b SI'.
Ho anche visto che, se era già presente il DSN, aggiornare semplicemente il Native Client non era sufficiente e falliva ugualmente l'installazione del vCenter: necessario cancellare e ricreare un nuovo System DSN. Questa cosa temo possa essere critica quando si fanno aggiornamenti: forse si deve creare un nuovo DSN che punti al db esistente e, quando si aggiorna, si seleziona quello.
Inoltre quando installate il VUM ricordatevi che, se non avete ancora aggiunto gli utenti come amministratori del vcenter (loggandosi nel webclient), dovete usare per forza l'utente administrator@vsphere.local. Anche se inserite l'utente Windows con cui gira il servizio vCenter prendete picche.
In Applications....I see userenv 1090 and VSS 8193, neither are referencing driver errors.
In System...I see Browser 8032 and Print 33. This is a print server.
Hi
are you doing local mode? What exact parameters are you looking to clarify? I will be happy to help.
-Nachiket
Ciao Rocco,
il post di Webster è completamente corretto: vCloud verrà dismesso per i clienti enterprise, mentre rimarrà per tutti i Service Provider con licenze VSPP. Per i primi, la soluzione di provisioning sarà vCAC, anche se immagino ci vorranno diverse versioni di entrambi per consentire un passaggio indolore, anche se temo il vero salto ci potrà essere solo quando VMware darà chiare indicazioni e procedure (e penso anche qualche tool) per realizzare le migrazioni. Inoltre, dovranno capire come travasare in vCAC alcuni dei costrutti che ad oggi sono in vCloud, come ad esempio il virtual networking. Probabilmente facendo parlare maggiormente vCAC con VCNS….
Per quanto riguarda gli altri tool, non ne conosco tanto, ho solo un’esperienza riportata da amici che lavorano in un provider che ha scelto come soluzione CIAC di Cisco (Cisco Intelligent Automation for Cloud, http://www.cisco.com/en/US/products/ps11869/index.html) più perchè sono uno shop Cisco che altro: networking e tutti server UCS già presenti. Per farla breve, è un bagno di sangue da quanto mi dicono, e la descrizione mi suona familiare rispetto a quanto sento anche per Openstack: il prodotto ha tantissime funzioni, molto promettente, molto potente, ma un casino da tirare in piedi, si spacca solo a guardarlo, e serve obbligatoriamente qualcuno del vendor per farlo andare.
Rispetto a questo, Openstack ha l’ulteriore problema di non avere un prodotto univoco, e infatti molte delle società che ci hanno messo il naso dentro ne hanno già realizzato versioni proprietarie, come HP ad esempio. Quindi rischi di ritrovarti con “quella” versione di Openstack, e addio a tutti i concetti di OpenSource e via dicendo. Openstack rischia veramente di finire in un mare di fork che poco si parlano l’un l’altro.
So infine che esiste qualcosa di simile di BMC, ma per esperienza con altri loro prodotti ne sto alla larga...
Io per ora resto con vCloud. Openstack se lo può permettere solo chi ha sufficienti sviluppatori interni per adattarlo e “farlo andare”.
Ciao,
Luca.
We've got that patch installed to fix the "black lines" issue, which the patch did fix. this issue is still happening.
however we narrowed it down to happening most frequently with Aero + Adjust for Best Appearing. Changing settings to Adjust for Best Performance significantly reduced the frequency of the issue, although it does still happen. We also found it happens via a Console connection, even after uninstalling the View Agent and reinstalling the Vmware Tools. so whatever the cause it seems to be with the VMs or HWv9 or maybe the ESXi hosts, similar to the "black lines"
...-DiskGb $_.HD1,$_.HD2...
Following Luc's suggestion above.
I am trying to selectively add a 2nd disk to some VMs only, so I had added 2 columns in my CSV and for the ones I don't need a second disk if I set to 0 the script fails for obvious reasons because of null value. I think this is the same issue as you mentioned about needing to change the code.
It works if I set the value of 2nd disk to 1 but I don't want to end up create additional disks for each VM unnecessarily.
Any ideas as to how I could simply go do this in my script to cater for this?
Thanks
Message was edited by: Mayur J Patel
Hi laurentsd,
By works on Windows, I meant works on browsers on Windows. Barring the exception that happens every now and then, here is the current story => It appears that there is a new filter in the web client that was absent on the desktop client i.e. Filtering by "My tasks" and "All users tasks". Under "My tasks" our posted tasks were NOT showing up and it was showing up in the "All users tasks". Also, these tasks have "Initiated By" set to nothing as opposed to native VMWare tasks that have the user information showing up in the "Initiated By". This setting was not present and not necessary earlier on the desktop client. I'll see if we are missing some setting that is required for the web client to publish this task under "My tasks".
Thank you
You could use a variable to pass to the parameter
$disks = @($_.HD1)
if($_.HD2 -ne 0){
$disks += $_.HD2
}
.... -DiskGB $disks...
I have the same problem on a number of HP BL460c G7. I have not tried this on the BL460c Gen8s, which use the same driver/firmware. My scenario right now is:
8 BL460c Gen8 running ESXi 5.1 with dvs Cisco Nexus 1000V. Trying to add 6 BL460c G7 with ESXi 5.5 to this dvs will lead to unrecoverable error in the elxnet driver. Adding the G7s to a dvs which is not Cisco based does not lead to the unrecoverable error.
I have engaged Emulex and they are currently trying to reproduce it.
Openstack se lo può permettere solo chi ha sufficienti sviluppatori interni per adattarlo e “farlo andare”.
questa è l'opinione che mi ero fatto su Openstack: allo spargersi di notizie (non del tutto fondate tra l'altro) "Ebay abbandona VMware per passare a Openstack" io replicavo che Ebay si può permettere un mezzo esercito di sistemisti e programmatori che lavorano *solo* su questa infrastruttura, non necessariamente quello che va bene per loro va bene anche per noi dove il povero cristo deve occuparsi dell'infrastruttura virtuale mentre sta cercando di scrivere un programmino a uso interno e contemporaneamente arriva l'utente a lamentare un problema al suo smartphone.
Anzi: probabilmente quello che va bene per Ebay non va bene per noi!
When you create a custom task on vCenter server, ensure you use vCenter username and not the NGC (SSO) username.
How are you creating the custom tasks on vCenter server?
I know sounds silly - but double check quotas (total and running vm limits) - just to make sure not an issue.
I know I am missing something but at a loss right now - might need to open case with VMware and get them the vcenter and more importantly VCD debug logs.
Thanks Luc for your fast reply.
Not being developer minded i couldn't think of using a variable. I just tried it and it works perfectly.
I was also looking to do something similar for adding additional NICs to selective VMs. Now that you have given me the idea I will do the same for that also.
Once again many thanks!
We are having an issue with our virtual machines losing access to, or suffering severe lag to, a mapped drive. It seems very random and we have no lead as to whether this is software or hardware related. I'll give a quick breakdown of the hardware and the config
SuperMicro X9DRH-7TF motherboard w/LSI 2208 SAS/SATA RAID controller onboard
96GB RDIMM 1600
2-Xeon E5-2690 processors
1-SanDisk Extreme SSD 240GB (attached to non RAID SATA port)-for OS (C:)
6-SanDisk Extreme SSD 480GB (attached to LSI controller-RAID 5)-for data (X:)
Host OS-Windows 7 Pro
24 VMs, using VMware Player Plus
Guest OS 1-5: Windows 7
Guest OS 6-24: Windows 8.1
Each VM has 3GB RAM allocated
The VMs all reside on the X drive and have a mapped network drive to X as they all need access to the same data set. We are running numerical modeling simulation software on each VM. Every once in awhile the VMs cannot access the X drive, or access is very slow. When this is happening even the host OS has slow access to the drive. We have had some issues with the Win8.1 VMs keeping a temporary lock on some of the files, which goes away after a few hours, not allowing moving or renaming. We didn't have the mapped drive issues when we tested with 24 VMs all running Win7. Anyone else had similar issues? Is this related Windows 8??
Yes, you should use createTask method of the WSSDK for creating tasks on vCenter Server. See the samples that are delivered with Web Client SDK for how to call vSphere Web services from vSphere web client.
One thing to remember that when you create a custom task on vCenter server, use vCenter username and not the NGC/SSO username. If you use NGC username then the progress of the task will not appear in
my recent tasks pane.
I am using free esxi 5.5, and no problem with 1TB usb disk, but disks of 2TB are not recognized.
Is there any known problem or restricition?
Thanks
I've found several bugs over the last few months that are really frustrating when trying to run Bash scripts using Invoke-VMScript. The work around has been to re-write the script or simply put them in a file, Copy-VMGuestFile, dos2unix and chmod +x and run it.
Here's a list of the bugs, I would appreciate if VMware would pay attention to these because I'm ready to give up on PowerCLI and move to a Ruby implementation of the SOAP API.
( exit 1 ); echo $?
This will return 0 from Invoke-VMScript, when it should return the value given to exit in the sub-script (1 in this case). When I pasted this to VMWare community forum, I was told to use the “ExitCode” Object returned by Invoke-VMScript. But there are more bugs.
[[ $(/bin/true) ]] || echo TRUE
This should return TRUE always, but when run from Invoke-VMScript we get:
bash: -c: line 0: syntax error near `]]’
bash: -c: line 0: `[[ ]] || echo TRUE’
Simple assigning a command output to a variable and echoing it.
A=$(grep alias ~/.bashrc |tail -1); echo $A
This returns nothing. Running ‘ps -ef’ while it’s executing reveals:
/bin/bash -c bash > /tmp/vmware-root/powerclivmware255 2>&1 -c “A=alias mv=’mv -i’; echo ; sleep 30”
Finally $! is broken. As an environmental variable in bash it should return the PID of the last process spawned when using &:
/bin/true & echo $!
Returns nothing.
I use the following Invoke-VMScript syntax:
Invoke-VMScript -VM "VM_Name" -ScriptType bash -ScriptText "blah" -GuestUser "root" -GuestPassword "blah"
It appears anywhere you use variables is likely to result in unexpected output. I'm escaping the $ symbol from within PowerShell and have even tried using here documents (@' <script> '@) to eliminate PowerShell as the cause. I've tested all of these scenarios in bash from the command line and all work as described.
EDIT: Forgot to mention using backticks instead of $() to get the output of an command yields the same results in examples 2 & 3
Luc, I just remembered there was one more thing I wanted your input on. Once my script creates the blank VMs I want to power on VMs in a particular order i.e.
1. DC
2. VC etc...
I am using the below method to power ON the DC VM:
#This line checks the powerstate and POWER ON Lan_DC VM to start installation
$vmState = Get-VM Lab_DC | Select-Object PowerState
If ($vmState -match "ff") {Get-VM Lab_DC | Start-VM}
but the unattended build can take about 30 mins to complete and this timing is not accurate because of my SATA disks (not the fastest). So is there a method I could employ for detecting the heartbeat of VMTools having started? In my DC build I install the VMtools as the last item and let it reboot. This status would indicate that the machine has been built and it is time for the VC VM to start up and start installing.