PCD 11.5.3 migration CUCM 9.1 to 11.5 fails 'unable to mount the floppy image'

[jstrittmatter]

Hello,

I'm trying to migrate a cluster with two CUCM 9.1.2.13900 and two IM&P 9.1.1.31900 server both to 11.5.1.11900.

PCD version is 11.5.3.10000 running on ESXi 5.5.

the UC cluster is discovered successfully and I have started a migration task to migrate all servers to new version. The task starts, successfully exports configuration from all nodes in the cluster.

But when it comes to the point where CUCM publisher is beeing installed, the taks failse with an error 'Unable to mount the floppy image named floppy/XXX_{publisher ip address}.flp stored in datastore PCD_NFS'

PCD log file contains error message like this:

2016-10-11 19:14:00,013 ERROR [DefaultQuartzScheduler_Worker-4] scheduler.InstallJob.configureVirtualMachine - The floppy image named floppy/1955_10.140.122.20.flp stored on datastore PCD_NFS could not be mounted by the VM named JB-BAW-CUCM1-NEW
com.cisco.ucmap.scheduler.SchedulerException: install.error.mount.floppy
    at com.cisco.ucmap.scheduler.InstallJob.configureVirtualMachine(InstallJob.java:243)
    at com.cisco.ucmap.scheduler.InstallJob.executeJob(InstallJob.java:469)
    at com.cisco.ucmap.scheduler.CustomJob.execute(CustomJob.java:215)
    at org.quartz.core.JobRunShell.run(JobRunShell.java:216)
    at org.quartz.simpl.SimpleThreadPool$WorkerThread.run(SimpleThreadPool.java:549)
2

When I use FileZilla to connect to PCD server, I can see that the directory /fresh_install/floppy that is mounted as PCD_NFS is empty. There is no file in this directory that could be mounted to the VM.

Any help is appreciated

Juergen

[asherif82]

Hello,

I am experiencing the same problem exactly. I have PCD version pcd_vApp_UCOS_11.5.3.10000-14_vmv7_v1.2 install and trying to install Call Manager 11.5.1.11900-26. However, it gave me below error:

"Unable to mount the floppy image named floppy/559_X.X.X.X.flp stored in datastore XXX_NFS onto the VM named XXX.Please check the ESXI logs for further details of the failure"

The only difference is that I have ESXi 6.0 installed on UCS servers C240M4.

Please advise.

[jstrittmatter]

Hello,

some further inverstigations showed that there is a relation to the ISO image you are using to install CallManager:

- upgrade only ISO images, that were made bootable i.e. with MagicISO are causing the error

- original bootable ISO image from cisco e-delivery or ordered using PUT seem to work

please tell me if this also solves your problem.

Juergen

[fhk-cseifert]

Hi Jürgen,

I struggled at the same point of migration of CUCM 9.1 to latest 11.5 with pcd. The export of all cucm nodes finished with no error, then the new vm powered on but nothing occurs. Neither floppy config nor bootable ISO mounted to the vm so migration task ran into timeout and failed.

I am using original bootable iso files and ova templates. We have esx 6.0 with enterprise license.  Boot device order of new vm is okay (CD-ROM first). Acl and routing is okay. 

I see the same error events at vsphere. There comes the command to mount the ISO and floppy, but there are no ISO and floppy at cpcd_NFS to mount. Should pcd server should not copy the installation ISO to nfs storage before? 

Now, I opened a tac case and Cisco engineer has no idea at first troubleshooting.

Did you have any new infos?

Thanks a lot in advance.

Greetings, Chris

[jstrittmatter]

Hello Chris,

root cuse in my case was wron ISO image. it was made bootable from an CCO downloaded upgrade image. PCD 11.5 for mistake detected this as an UNRESTRICTED image.Here are the relevant lines from the ucmap log file.

You have to use original bootable ISO image.

Jürgen

2016-10-12 00:30:22,728 INFO  [DefaultQuartzScheduler_Worker-9] scheduler.InstallJob.getStandardIsoFileName - Finding the exact filename for the selected iso : /common/adminsftp/fresh_install/UCSInstall_UCOS_11.5.1.11900-26.sgn.iso

2016-10-12 00:30:34,527 INFO  [DefaultQuartzScheduler_Worker-9] ucmap.CommandExecute.executeCommand - The iso name is :     local unrestricted_product_prefix="UCSInstall_UCOS_UNRST_"

2016-10-12 00:30:34,527 INFO  [DefaultQuartzScheduler_Worker-9] scheduler.InstallJob.getStandardIsoFileName - The proper version from thge selected iso is :     local unrestricted_product_prefix="UCSInstall_UCOS_UNRST_"

2016-10-12 00:30:34,527 INFO  [DefaultQuartzScheduler_Worker-9] custom.ApplicationCustom.getVersionNumber - file name     local unrestricted_product_prefix="UCSInstall_UCOS_UNRST_"

2016-10-12 00:30:34,527 INFO  [DefaultQuartzScheduler_Worker-9] custom.ApplicationCustom.getVersionNumber - version is      =""

2016-10-12 00:30:34,527 INFO  [DefaultQuartzScheduler_Worker-9] custom.ApplicationCustom.getVersionNumber - digits size 1

2016-10-12 00:30:34,528 INFO  [DefaultQuartzScheduler_Worker-9] scheduler.InstallJob.getStandardIsoFileName - Finding the exact filename for the selected iso : /common/adminsftp/fresh_install/UCSInstall_UCOS_11.5.1.11900-26.sgn.iso

2016-10-12 00:30:41,559 INFO  [DefaultQuartzScheduler_Worker-9] ucmap.CommandExecute.executeCommand - The iso name is :     local unrestricted_product_prefix="UCSInstall_UCOS_UNRST_"

2016-10-12 00:30:41,559 INFO  [DefaultQuartzScheduler_Worker-9] scheduler.InstallJob.getStandardIsoFileName - The proper version from thge selected iso is :     local unrestricted_product_prefix="UCSInstall_UCOS_UNRST_"

2016-10-12 00:30:41,670 ERROR [DefaultQuartzScheduler_Worker-9] inventory.M1Install.createMigrationFlp - Failed to get from or to NodeInfo

java.lang.StringIndexOutOfBoundsException: String index out of range: -62

            at java.lang.String.substring(String.java:1911)

[panjwani80]

Chris,

How did you fix your issue? I can't get the ISO mounted to ESXi datastore either with ESXi 6.0 and PCD 11.5.3

Thanks

Aamir

[fhk-cseifert]

Hi Aamir,

did you already uploaded the bootable ISO to PCD and PCD says it is valid? If yes, then PCD will mount his own pcd_NFS datastore to esxi. At second step, PCD will mount the intial ucm-config (with exported ip, hostname, ...) as floppy image for installation process.

That was one of my fault, that I built an own nfs-store which PCD should use, but he has to do it self (with esxi root password).

Hope this help.

Greetings, Chris

[asherif82]

Hello Juergen,

I managed to solve that problem. In my case I was having an issue in the floppy attached with the OVA template as it was somehow not working. I removed it and add it again. In addition, I changed the IOS image of CUCM with a properly bootable one and it worked like magic.

Regards,

Amr Sherif