Support Questions

Find answers, ask questions, and share your expertise
Announcements
Celebrating as our community reaches 100,000 members! Thank you!

Problem distributing CDH parcel to manager host

avatar
Explorer

Experiencing another weird error.

 

I've created an EC2 instance for Cloudera Manager, and went to it's webconsole.  I was presented with add hosts screen, which recommended i added the manager host, which I did.  However, upon distribution of parcels, i am getting a repeated error untarring CDH 5.1.0.  When I say repeated, it is essentially stuck in a loop trying to distribute, failing, then restarting distributing again.

 

thanks in advance if you've already seen this!

 

error.png

 

 

1 ACCEPTED SOLUTION

avatar
This can also happen if you have insufficient disk space in /opt/cloudera/

As Gautam said, agent logs should indicate the error.

View solution in original post

7 REPLIES 7

avatar

So is this a single host cluster? To troubleshoot parcel distribution, check the directories under /opt/cloudera. There could be a dangling symlink in one of those directories.

 

# ls -l /opt/cloudera/parcel*

 

Also check the agent logs where you'll see the real cause of failure. This is under /var/log/cloudera-scm-agent/*.log. If you can, please paste the log snippet here (or use a service like pastebin or github gists)

 

Regards
Gautam

 

Regards,
Gautam Gopalakrishnan

avatar
Explorer

thank you for the hints.

 

the parcel directory looks ok to me, although two of the directories are owned by root rather than cloudera-scm.  however, making that in a symbolic link made the problem go away.  thanks!

 

root@ip-10-133-2-15:/opt# ls -l /opt/cloudera/parcel*
/opt/cloudera/parcel-cache:
total 4
drwx------ 3 root root 4096 Aug 14 00:10 tmpYThWAd

/opt/cloudera/parcel-repo:
total 1687240
-rw-r----- 1 cloudera-scm cloudera-scm 1726036186 Aug 13 21:47 CDH-5.1.0-1.cdh5.1.0.p0.53-precise.parcel
-rw-r----- 1 cloudera-scm cloudera-scm 41 Aug 13 21:47 CDH-5.1.0-1.cdh5.1.0.p0.53-precise.parcel.sha

/opt/cloudera/parcels:
total 0

 

 

avatar

The permissions look fine. The temp file though looks like it's a leftover of a previous aborted download. It's possible your network is slow and the downloads are timing out. You should delete it.

 

What do the agent logs say? That's where you'll see the real cause of problems.

 

 

Regards,
Gautam Gopalakrishnan

avatar
This can also happen if you have insufficient disk space in /opt/cloudera/

As Gautam said, agent logs should indicate the error.

avatar
Explorer

sadly, the logs for that particular day have been blown away.  the dangling symlink was the issue, as it was not pointing to the bigger storage device, and root store was probably getting full in the middle of the operation.

 

thanks for your help!

avatar
Contributor

i have same problem. can u help what u did to resolve this

avatar
New Contributor

i got the same issue,

agent logs:

flood_seed_timeout': 0, u'eventserver_port': None}
Traceback (most recent call last):
File "/opt/cloudera/cm-agent/lib/python2.7/site-packages/cmf/agent.py", line 1568, in handle_heartbeat_response
self._handle_heartbeat_response(response)
File "/opt/cloudera/cm-agent/lib/python2.7/site-packages/cmf/agent.py", line 1730, in _handle_heartbeat_response
self._update_parcel_activation_state(response)
File "/opt/cloudera/cm-agent/lib/python2.7/site-packages/cmf/agent.py", line 1641, in _update_parcel_activation_state
manage_new_parcels)
File "/opt/cloudera/cm-agent/lib/python2.7/site-packages/cmf/parcel.py", line 624, in configure_all_symlinks
prod = self[name]
File "/opt/cloudera/cm-agent/lib/python2.7/site-packages/cmf/parcel.py", line 798, in __getitem__
return self.products.__getitem__(key)
KeyError: 'CDH'