Member since
06-26-2015
515
Posts
140
Kudos Received
114
Solutions
My Accepted Solutions
| Title | Views | Posted |
|---|---|---|
| 2570 | 09-20-2022 03:33 PM | |
| 6937 | 09-19-2022 04:47 PM | |
| 3665 | 09-11-2022 05:01 PM | |
| 4274 | 09-06-2022 02:23 PM | |
| 6777 | 09-06-2022 04:30 AM |
02-12-2022
08:02 PM
1 Kudo
If the different types of files are in different directories in HDFS, for example, you can use Expression Language to set the values for fragment.index and metric, using a single ListHDFS -> FetchHDFS -> UpdateAttribute. The expression below sets the value for metric according to the path where the file came from: ${path:equals("/tmp/input/dir1"):ifElse("a", ${path:equals("/tmp/input/dir2"):ifElse("b", ${path:equals("/tmp/input/dir3"):ifElse("c", ${path:equals("/tmp/input/dir4"):ifElse("d", ${path:equals("/tmp/input/dir5"):ifElse("e", ${path:equals("/tmp/input/dir6"):ifElse("f", "other")})})})})})} You can do the same for fragment.index.
... View more
02-12-2022
07:40 PM
How do you differentiate the files in HDFS? Are they in different directories? Have different filenames?
... View more
02-12-2022
05:30 PM
Are all the files similar and you assign the fragment indexes in a round robin fashion? (1,2,3,4,5,6,1,2,3,...) Or do the different index numbers identify different types of files? When you merge, can you merge as many files as possible or do they always need to be merged 6 by 6? Can you give an example of how you are going to use the index in the QueryRecord processor?
... View more
02-12-2022
05:21 PM
From the extension of your key file (key.ppk), my guess is that you're using PuTTY to connect to the VMs. Is that correct? PuTTY uses a different key format than OpenSSH clients. If the above is correct, try converting your key.ppk to OpenSSH format using PuTTYgen (see link below) and try again using the converted file. https://www.thegeekdiary.com/how-to-convert-puttys-private-key-ppk-to-ssh-key/ Cheers, André
... View more
02-12-2022
11:03 AM
Which steps did you take to enable Kerberos? Did you use the wizard in Cloudera Manager? How many nodes does your cluster have? Which node are you running these commands from? Have you tried from others nodes (e.g have you tried from the Name node host?)
... View more
02-12-2022
05:06 AM
Try changing udp_preference_limit to 1 in the krb5.conf file on all the hosts and restart your cluster. Also notice that you have a typo in that parameter's name. The correct is udp_preference_limit and not udp_prefrence_limit.
... View more
02-11-2022
01:51 PM
1 Kudo
Great to hear! I try my best to understand Jolt because sometimes it can be quite useful, but I think it has a very convoluted syntax and sometimes it's really hard to use. But practice helps. The first asterisk matches against the field names of an object. The second asterisk depends: if the value of the attribute is a scalar, it will match against the value; if it's a nest object, it will match against the name of the nested object. The trick is that when it matches the value of the object it does not match nulls 😉 Cheers, André
... View more
02-11-2022
04:32 AM
This looks like your HDFS service is misconfigured. Are you using CDP or open-source HDFS? Could you please share your HDFS configuration, specifically the properties that you set to enable Kerberos? André
... View more
02-11-2022
01:40 AM
Thanks. The context of it is much clearer now. Could you also provide a screenshot of the Login Credentials page of the wizard?
... View more