Riak doesn't start anymore after server crash - no errors logged

Steven Sylvester ssylvester at basho.com
Wed Nov 11 19:05:53 EST 2015


Hi Markus,

One thing that stands out in what you pasted is the following line:

2015-11-09 11:31:45.263 [info] <0.73.0> alarm_handler:
{set,{{disk_almost_full,"/"},[]}}

Check the output of `df` and add nodes/disk as necessary to keep your disk
usage at a manageable level. Our docs recommend keeping disk use below 60%:
http://docs.basho.com/riak/latest/ops/building/planning/cluster/#Disk

Disk capacity issues can sneak up on you due to some backend operations
that require a lot of space all at once, such as Bitcask merging or AAE.
*Bad things* tend to happen to a Riak cluster with full disks, including
data loss/corruption.

Other than that, are there any changes that have been made to the
environment lately? Have you performed any upgrades recently?

Check the output of `ps aux` and kill any processes related to Riak, such
as any instances of `beam.smp` before attempting to start the node again.

If the node won't start after these troubleshooting steps, run `riak
console`  which will attempt to start the node and put you into an Erlang
attach shell. Monitor the messages that are printed to stdout and if/when
it seems to hang, paste the output of the whole thing here and we can take
a look.

-- 
Steven Sylvester, CISSP
Client Services Engineer
Basho Technologies
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.basho.com/pipermail/riak-users_lists.basho.com/attachments/20151111/9cc847ae/attachment-0002.html>


More information about the riak-users mailing list