You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@storm.apache.org by sara smith <qd...@gmail.com> on 2018/01/16 10:30:39 UTC
Help me with this issue
Which directories should the user running Apache Storm be granted access?
Storm path is /opt/storm
user/group: storm:storm
Storm 1.0.3 on Red Hat Enterprise Linux 7.3.
Already did:
sudo chown -R storm:storm /opt/storm/
sudo chmod -R 775 /opt/storm
3 machines .. 1 running nimbus and 2 running supervisors. all services
start fine without errors on logs.. the error described on the previous
messages happens specifically when trying to submit the topology jar.
Unfortunately, this error continues for days) even after terminating the
instances and rebuilding instances from scratch and it is now preventing
from going live with the MVP in production.
Any help will be appreciated.
--
Dani
+1-9493428085
Business Information Network
Co-Founder
https://www.binmy.com
Re: Help me with this issue
Posted by Øystein Larsen <oy...@ifi.uio.no>.
On 16/01/2018 15:12, sara smith wrote:
> I will try strace, but a frind of mine told me to edit Storm.yaml, and
> check disabled ACL for blob. I tried from /tmp but I still get the same
> error .
Could you post your storm.yaml file here? Sanitize as required.
Re: Help me with this issue
Posted by sara smith <qd...@gmail.com>.
I will try strace, but a frind of mine told me to edit Storm.yaml, and
check disabled ACL for blob. I tried from /tmp but I still get the same
error .
On Tue, Jan 16, 2018 at 3:50 PM, Øystein Larsen <oy...@ifi.uio.no> wrote:
> On 16/01/2018 11:30, sara smith wrote:
> > Which directories should the user running Apache Storm be granted
> > access? Storm path is /opt/storm user/group: storm:storm Storm 1.0.3
> > on Red Hat Enterprise Linux 7.3. Already did: sudo chown -R
> > storm:storm /opt/storm/ sudo chmod -R 775 /opt/storm 3 machines ..
> > 1 running nimbus and 2 running supervisors. all services start fine
> > without errors on logs.. the error described on the previous
> > messages happens specifically when trying to submit the topology jar.
> > Unfortunately, this error continues for days) even after
> > terminating the instances and rebuilding instances from scratch and
> > it is now preventing from going live with the MVP in production. Any
> > help will be appreciated.
> >
>
> - The storm home directory, recursively
> - The storm log directory
> - The directory structure holding the topology code
>
> Do you use Streamparse for deploying the topology?
>
> Are you running the topology on hosts where a MAC like SElinux or
> AppArmor may block processes from accessing the required files and
> directories?
>
> If you try running the topology on the Nimbus host alone (which is in
> any case an excellent way for error-checking) you could try to pass it
> through strace, tracing specifically file-related syscalls. That trick
> will usually give a failing process away, when you search for DENIED in
> the trace output. Yes, it is crude, but so is existence.
>
--
Dani
+1-9493428085
Business Information Network
Co-Founder
https://www.binmy.com
Re: Help me with this issue
Posted by Øystein Larsen <oy...@ifi.uio.no>.
On 16/01/2018 11:30, sara smith wrote:
> Which directories should the user running Apache Storm be granted
> access? Storm path is /opt/storm user/group: storm:storm Storm 1.0.3
> on Red Hat Enterprise Linux 7.3. Already did: sudo chown -R
> storm:storm /opt/storm/ sudo chmod -R 775 /opt/storm 3 machines ..
> 1 running nimbus and 2 running supervisors. all services start fine
> without errors on logs.. the error described on the previous
> messages happens specifically when trying to submit the topology jar.
> Unfortunately, this error continues for days) even after
> terminating the instances and rebuilding instances from scratch and
> it is now preventing from going live with the MVP in production. Any
> help will be appreciated.
>
- The storm home directory, recursively
- The storm log directory
- The directory structure holding the topology code
Do you use Streamparse for deploying the topology?
Are you running the topology on hosts where a MAC like SElinux or
AppArmor may block processes from accessing the required files and
directories?
If you try running the topology on the Nimbus host alone (which is in
any case an excellent way for error-checking) you could try to pass it
through strace, tracing specifically file-related syscalls. That trick
will usually give a failing process away, when you search for DENIED in
the trace output. Yes, it is crude, but so is existence.