Conversation
Update quickstart.md to contain explicitly all the necessary commands needed to setup the accumulo environment on a single node. When starting only tserver, some of the commands (e.g. : `createtable mytable` )executed on the accumulo shell will hang undefinitely
|
|
||
| accumulo-service master start | ||
| accumulo-service tserver start | ||
| accumulo-service monitor start |
There was a problem hiding this comment.
You can use the accumulo-cluster cmd to start all of the services either on a single node or on multiple nodes - these commands are only necessary if you elect to run each manually (or say you need to restart just one) Maybe it would be more user-friendly if the accumulo-cluster command section came before this section?
Also, for a complete installation I believe that you might be missing necessary services (gc). Then maybe this could be
accumulo [service-name] or accumulo-service [service-name] start | stop with a more complete list of the required service names?
There was a problem hiding this comment.
@EdColeman i'm a newbie with accumulo and while trying to create a table i simply didn't know why it hangs.
I've spent more than an hour debugging and pondering about it, and afterwards I came to see the fact that there are more services to start, not only the tserver.
Obviously my modification is quite naive because there are probably more services. So what you are suggesting with accumulo [service-name] fits much better.
The quick start for me wasn't so quick :)
I came across the sample code https://github.com/apache/accumulo-examples/blob/main/docs/sample.md which gives a few pointers on how to work with the accumulo shell - I find it quite good for a quick start - to get a feeling about what accumulo actually does.
There was a problem hiding this comment.
The docs start with consider using fluo-uno - that's the easiest on ramp that I know of. Is there a reason that you did not try it? Would changing / adding wording there have made you more likely to use fluo-uno?
Moving the cluster command before the individual commands might have made your experience easier - with accumulo-cluster start, things should have been started and removes the beginner from needing five commands instead of one.
But with that, uno fetch accumulo, uno start accumulo is way easier than setting up accumulo from scratch.
There was a problem hiding this comment.
A bit unrelated, but another small issue I've encountered was the ClassNotFoundError regarding zookeeper's KeeperException class.
Cause of it :
accumulo-env.sh
CLASSPATH="${CLASSPATH}:${lib}/*:${HADOOP_CONF_DIR}:${ZOOKEEPER_HOME}/*:${HADOOP_HOME}/share/hadoop/client/*"
replaced it with
CLASSPATH="${CLASSPATH}:${lib}/*:${HADOOP_CONF_DIR}:${ZOOKEEPER_HOME}/lib/*:${HADOOP_HOME}/share/hadoop/client/*"
(added lib after ZOOKEEPER_HOME)
There was a problem hiding this comment.
But with that, uno fetch accumulo, uno start accumulo is way easier than setting up accumulo from scratch.
I'm looking now through the code of uno and see that it downloads everything what is needed. I am sorry, but due to the fact that I have already zookeeper and hadoop on my machine I thought to opt to simply install the bin archive of accumulo.
There was a problem hiding this comment.
For the zookeeper issue we'd need to know the zookeeper and accumlo versions. There were changes in ZooKeeper 3.x series that modified where zookeeper store jars and separated the jute (zk comms layer) into a separate jar. That's one benefit of uno - it should download compatible versions of things - it also allows you to point at you local reop and build / run accumulo with your changes if you think you might go down that route in the future.
There was a problem hiding this comment.
I tried with both apache zookeeper 3.5.9 and zookeeper 3.7.0
In both of them, the .jar libraries are located under ZOOKEEPER_HOME/lib directory.
There was a problem hiding this comment.
What version of Accumulo?
There was a problem hiding this comment.
Since that is unrelated, let's not bog this issue down with that issue. See apache/accumulo#1530 for more.
| Start Accumulo processes (tserver, master, monitor, etc) using command below: | ||
|
|
||
| accumulo master | ||
| accumulo tserver | ||
| accumulo monitor |
There was a problem hiding this comment.
I think this was an example explaining how to start an individual service, generally. Rather than add the other services as an example, it would probably be better to reword the instructions to make it clear that this is just one example.
(similar comment below)
There was a problem hiding this comment.
I made some suggested improvements in #271 . Please take a look and see if they help alleviate the concerns you were trying to address in this PR.
There was a problem hiding this comment.
It looks good to me, but now I am bit biased because I read in the meantime more about the architecture of accumulo and know now a bit how to use it.
I'd highlight which of the services need to run on the local installation at the very least. I don't run for example GC and it's still fine. You may say that I should have used Uno, but maybe some other users will try their luck also by doing directly a local installation of accumulo.
Also important is a link to the wonderful accumulo-examples repository. This is a gem containing quite useful stuff to get started with accumulo.
Troubleshooting is also quite welcome. Without troubleshooting tips some of the users will stop early after trying to setup accumulo.
|
Latest
2.0.1
…On Thu, Apr 8, 2021 at 4:36 PM EdColeman ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In _docs-2/getting-started/quickstart.md
<#270 (comment)>
:
> accumulo-service tserver start
+ accumulo-service monitor start
What version of Accumulo?
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#270 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ANU3TAP2KO54YS3ZBONKBHTTHW5O3ANCNFSM42Q32WMA>
.
|
|
The pom for 2.0.1 specifies zookeeper 3.4.14. For the next release (main branch), the pom currently has ZooKeeper version 3.5.9 and it looks like the accumulo-env.sh has been updated with the zookeeper jar locations ($ZOOKEEPER_HOME/lib/) for later zookeeper releases. |
|
Thank you @EdColeman and @ctubbsii for taking the time to talk with me on the quick start documentation topic. |
3 participants
Update quickstart.md to contain explicitly all the necessary commands needed to setup the accumulo environment on a single node.
When starting only tserver, some of the commands (e.g. :
createtable mytable)executed on the accumulo shell will hang undefinitely