PST Locator fails to get the list of volumes from a NetApp vfiler

[Olivian_Retezan]

Hi *,

We're trying to implement PST Migration.
The Enterprise Vault is version 8 SP2.
The NetApp is a vfiler with ONTAP version 7.3.1.
The permission style on the vfiler's root volume is set to unix style (this can't be changed).
The EV service account is an administrator of the vfiler.

When running the PST Locator task against this filer, the following error is logged by the task:
134089 17:11:32.314 [588] (PstLocatorTask) <5152> EV:M CBaseDirectoryServiceWrapper::CreateDirectoryService() - Entry [m_nNumTries = 40]
134090 17:11:32.330 [588] (PstLocatorTask) <5152> EV:L CBaseDirectoryServiceWrapper::CreateDirectoryService() - Successfully communicated with an EV Directory Service on the local machine
134091 17:11:32.392 [588] (PstLocatorTask) <5152> EV:M CNetAppShareLister::GetShareList - Entry with filer [NetApp_vfiler_name]
134092 17:11:33.361 [588] (PstLocatorTask) <5152> EV:M CNetAppShareLister::GetShareList - Error list-iter-start 13003: User domain\service_account does not have capability to invoke API cifs-share-list-iter-start.
134093 17:11:33.361 [588] (PstLocatorTask) <5152> EV:M CNetAppShareLister::GetShareList - Caught HRESULT [0x80004005]
134094 17:11:33.361 [588] (PstLocatorTask) <LocateHardDiskThread:5152> EV-H {PstLogFile.LogError} Hard disk search failure on NetApp_vfiler_name: Failed to get list of volumes on NetApp Filer : Error HRESULT E_FAIL has been returned from a call to a COM component.

Within the Admin console, on the computers section, the error "Failed to get list of volumes on NetApp Filer : Error HRESULT E_FAIL has been returned from a call to a COM component." is listed on the hard disk search status section. The "Last Hard Disk Scan" is not filled.

On the compatibility list, the configuration we have is considered as supported.

Do you have an idea on how to fix this?
Thanks in advance!

[Maverik]

Have you got a dtace showing this at all? If so can you post it.

[Olivian_Retezan]

My original post contains the information logged by the PSTLocator task in the dtrace.

[Rob_Wilcox1 · Accepted solution]

So my question would be to NetApp...

Is cifs-share-list-iter-start "allowed" when the root volume is Unix style?   If it's allowed are permissions required explicitly to be able to call that API?

I can see in our code that we call that function and the error we are getting back is the 13003 one from your trace. 

[Maverik]

Apologies, must of gone blind at that time. Sorry dont know what this is.

[Maverik]

Apologies, must of gone blind at that time. Sorry dont know what this is.

[Olivian_Retezan]

Hi Rob,

I have spoken with the storage guys and they added the EV service account to a new role that has, among other, the capability cli-cifs* allowed.

After this, I started again the PSTLocator task and it seems it started to list the filer volumes on the logs. The task has not finished yet, but it looks really promicing.

I'll wait a bit more, so the task is able to finish and come back with the results.
Thanks a lot for the tip!
Br, Olivian

[Olivian_Retezan]

No worries.

[Rob_Wilcox1]

You're welcome...  I am interested in the outcome.

[Olivian_Retezan]

Hi Rob,

It actually worked! Now the PST Locator was able to list all the PSTs from the vfiler. I'll test the same scenario on the real environment and continue with the migration.

Thanks a lot on this!
Br, Olivian

[Rob_Wilcox1]

I'm glad it is working for you .. I'm not sure whether what you're seeing is a side effect of OnTap 7.3, or a side effect of having only Unix style permissions.  I've asked our certification folks to take a look at this.

[Olivian_Retezan]

Hi Rob,

On the productive environment we get the following error when running the PST Locator task:
45621 15:12:41.601 [6500] (PstLocatorTask) <LocateHardDiskThread:7148> EV-L {PstLogFile.LogMessage} Hard Disk search on productive_vfiler is starting...
45623 15:12:41.601 [6500] (PstLocatorTask) <7148> EV:M CBaseDirectoryServiceWrapper::CreateDirectoryService() - Entry [m_nNumTries = 40]
45624 15:12:41.601 [6500] (PstLocatorTask) <7148> EV:L CBaseDirectoryServiceWrapper::CreateDirectoryService() - Successfully communicated with an EV Directory Service on the local machine
45625 15:12:41.617 [6500] (PstLocatorTask) <7148> EV:M CNetAppShareLister::GetShareList - Entry with filer [productive_vfiler]
45626 15:12:43.867 [6500] (PstLocatorTask) <7148> EV:M CNetAppShareLister::GetShareList - Error list-iter-start 13011: RPC Error - The RPC server is unavailable.
45627 15:12:43.867 [6500] (PstLocatorTask) <7148> EV:M CNetAppShareLister::GetShareList - Caught HRESULT [0x80004005]
45628 15:12:43.867 [6500] (PstLocatorTask) <LocateHardDiskThread:7148> EV-H {PstLogFile.LogError} Hard disk search failure on productive_vfiler: Failed to get list of volumes on NetApp Filer : Error HRESULT E_FAIL has been returned from a call to a COM component.

The only difference from the test environment is that the EV server does not have any SP installed (8.0 in the productive environment instead of 8.0 SP2).
If I try to browse to the filer share I have read/write access with the EV service account.

Could this issue be fixed by SP2?

Thank you in advance!
Olivian

[Rob_Wilcox1]

Olivian,

I'm not aware of a fix to address this in 8 SP 2.   However if your network, server builds and so on are the same in your lab versus your production environment, apart from the EV Service Pack version, and because any future-fix-if-it-is-a-bug would be based on Service Pack 2, then maybe going to Service Pack 2 is the best thing.

Before doing that though, I'd review the environments for any other differences.

Remember we are getting that error back when we make the call to cifs-share-list-iter-start...  maybe it's possible to contact NetApp, and/or produce a small demo program which makes the same call to see if that works?

[Olivian_Retezan]

Thanks for the tips. I'll try to get in contact with the storage team again.