Hi, I've been trying to understand the smb-processor landscape lately, and have run into some problems.
1) Timeout It seems like timeout was introduced with the SmbjClientProviderService, and SmbUtils.java defines the withTimeout setting from the underlying smbj library. In SmbProperties.java, the default timeout value is set to 5 sec. The problem is that this code is used by the older GetSmbFile and PutSmbFile processors as well, but the timeout configuration item is not exposed in their config ui. 2) Minimum File Age The ListSmb processor supports "Minimum File Age", which is great, but the older GetSmbFile does not. GetSmbFile does not support incoming connections either, so you cannot use ListSmb before GetSmbFile to select files. 3) Keep Source File The older GetSmbFile processor has the "Keep Source File" setting for removing files after you get them, which is nice. However, the newer FetchSmb does not appear to support this. As a result, you cannot get and delete files with a minimum age greater than X. 4) DFS referrals The underlying hierynomus/smbj library supports DFS referrals with the withDfsEnabled settings. The default value for this in the library is False. It would be very useful if this setting was exposed in all smb related processors and services that use hierynomus/smbj. I created https://issues.apache.org/jira/browse/NIFI-12837 for that. I might misunderstand something here as I'm not a java coder. Tried to patch the DFS feature, and it started working against the cifs cluster in my environment at least. I also attempted to patch the timeout ui setting in PutSmbFile, and that seemed to resolve some of the timeout issues I was seeing with PutSmbFile as well. But I don't know if there is some overarching plan or goal when it comes to the smb processors, or should I just create jira tickets for these things? -- Anders Synstad
