Ignoring the prerequisite: "are HDFS, ZooKeeper and Accumulo all running
properly to the best of your knowledge"...
What does you table look like? How many rows? How many columns per row?
Are the number of columns per row fixed or varied? Do you have any
locality groups configured? Have you made any changes to the default
configuration (e.g. files per tablet, disabled compactions, etc)?
If you open another shell while running a scan on this column family
that "hangs", try running the `listscans` command. You should see the
scan from your other shell as well as the server it's trying to
communicate with. This should narrow down which server's logs to
inspect. On that server:
* check the OS load (is it actually doing something)
* check the logs for any exceptions/errors
* check memory usage for the TabletServer (the logs also contain regular
heap-size messages you can grep for).
If you're still stuck here, while you have this scan running, try
`jstack`'ing that TabletServer a few times in succession and redirect it
to a file. We should be able to piece together some sort of guess as to
what the TabletServer is doing.
- Josh
z11373 wrote:
Hi,
I just experienced weird Accumulo issue I never seen before.
Running 'scan -t TABLE1 -c foo' from the shell, it just hung.
However, running scan for another column family (still on same table), i.e.
'scan -t TABLE1 -c bar' returns immediately. I looked at Accumulo logs, and
can't really figure out what's wrong. I tried with third column family, and
it also returns immediately.
It looks like the problem is only for specific column family, is that even
possible??
Any idea on how to troubleshoot this?
Thanks,
Z
--
View this message in context:
http://apache-accumulo.1065345.n5.nabble.com/scan-command-hung-tp15286.html
Sent from the Developers mailing list archive at Nabble.com.
