Closed GoogleCodeExporter closed 8 years ago
Looks like you are using an old Java API with a newer Scalaris. We have changed
some APIs for 0.3.0 compared to previous releases.
Can you tell me which version of Scalaris you are using?
For svn trunk: which revision?
For rpm/deb packages: which packages and which versions?
Please also verify that the correct version is started in case you have
multiple versions installed.
Original comment by nico.kru...@googlemail.com
on 12 Dec 2011 at 8:23
I am using scalaris-0.3.0 by unpacking scalaris_0.3.0.orig.tar.gz downloaded
from http://download.opensuse.org/repositories/home:/scalaris/xUbuntu_10.04/.
I dont have multiple versions.
I am currently getting the following error.
java de.zib.scalaris.Main -w "rt" "4" -v
ConnectionFactory properties:
config file =
scalaris.node = [node1@domU-12-31-39-02-7E-DC.compute-1.internal]
scalaris.cookie = chocolate chip cookie
scalaris.client.name = java_client
scalaris.client.appendUUID = true
write failed with unknown:
de.zib.scalaris.UnknownException: Erlang message:
{badrpc,{'EXIT',{badarg,[{comm,send_local,2},{rdht_tx_read,work_phase,3},{rdht_t
x_write,work_phase,3},{rdht_tx,'-initiate_rdht_ops/1-lc$^0/1-0-',1},{rdht_tx,req
_list,2},{rdht_tx,req_list,2},{api_tx,req_list,2},{api_tx,write,2}]}}}
at de.zib.scalaris.CommonErlangObjects.processResult_commit(CommonErlangObjects.java:159)
at de.zib.scalaris.TransactionSingleOp.write(TransactionSingleOp.java:369)
at de.zib.scalaris.TransactionSingleOp.write(TransactionSingleOp.java:399)
at de.zib.scalaris.Main.main(Main.java:138)
Original comment by subhojit...@gmail.com
on 13 Dec 2011 at 12:14
I should have seen this earlier, but "scalarisctl restart -f" is not enough to
start scalaris as this will only start an empty erlang shell with some basic
services but _NO_ Scalaris node in it.
Use e.g. "scalarisctl restart -f -s".
For a complete list of options, refer to "scalarisctl -h".
If this did not help, consider the following:
Did you just extracted it and then "./configure && make"?
Or did you also install it with "make install"?
Please also check for any error messages during make (warnings should be ok;
they are mostly in third party libraries).
Please also run "scalarisctl checkinstallation" (either from ./bin or from
where you installed it to).
Original comment by nico.kru...@googlemail.com
on 14 Dec 2011 at 5:49
I tried scalarisctl checkinstallation and got:
INFO: could not find Scalaris' Java-API files. You won't be able to use the
'scalaris' command line script to access Scalaris.
'make java' will build the Java-API
We were trying to run: /usr/bin/scalaris
Yes I also did ./configure and make and got:
configure: arity() is available
{"init terminating in
do_boot",{undef,[{httpc,set_option,[max_sessions,1]},{init,start_it,1},{init,sta
rt_em,1}]}}
configure: httpc unavailable, using wrapper module
checking for /usr/share/java-utils/java-functions... no
configure: WARNING: java-functions was not found in
"/usr/share/java-utils/java-functions" (java-api/scalaris may fail)
checking for build-classpath... no
configure: WARNING: build-classpath was not found in "" or is not executable
(java-api/scalaris may fail)
Original comment by subhojit...@gmail.com
on 15 Dec 2011 at 12:00
did you try your test again starting scalaris with the following command?
scalarisctl restart -f -s
this is the most likely cause of the last error you reported.
And since you did not start the Java-API through our command line wrapper, the
warnings from checkinstallation and configure do not apply to you.
(we use the two tools to build the class path in our scalaris command line
wrapper - if they are not available, paths may be wrong)
Original comment by nico.kru...@googlemail.com
on 15 Dec 2011 at 9:44
Hi,
I tried scalarisctl restart -f -s. There was no error as before. But
running scalarisctl checkinstallation gave:
INFO: could not find Scalaris' Java-API files. You won't be able to
use the 'scalaris' command line script to access Scalaris.
'make java' will build the Java-API
And as before, /java-api/scalaris -w "rt" "4" -v gives
ConnectionFactory properties:
config file =
scalaris.node = [node1@domU-12-31-39-02-7E-DC.compute-1.internal]
scalaris.cookie = chocolate chip cookie
scalaris.client.name = java_client
scalaris.client.appendUUID = true
write failed with unknown:
de.zib.scalaris.UnknownException: Erlang message:
{badrpc,{'EXIT',{badarg,[{comm,send_local,2},{rdht_tx_read,work_phase,3},{rdht_t
x_write,work_phase,3},{rdht_tx,'-initiate_rdht_ops/1-lc$^0/1-0-',1},{rdht_tx,req
_list,2},{rdht_tx,req_list,2},{api_tx,req_list,2},{api_tx,write,2}]}}}
at de.zib.scalaris.CommonErlangObjects.processResult_commit(CommonErlangObjects.java:159)
at de.zib.scalaris.TransactionSingleOp.write(TransactionSingleOp.java:369)
at de.zib.scalaris.TransactionSingleOp.write(TransactionSingleOp.java:399)
at de.zib.scalaris.Main.main(Main.java:138)
Original comment by subhojit...@gmail.com
on 15 Dec 2011 at 9:54
Hi,
I tried scalarisctl restart -f -s. There was no error as before. But running
scalarisctl checkinstallation gave:
INFO: could not find Scalaris' Java-API files. You won't be able to use the
'scalaris' command line script to access Scalaris.
'make java' will build the Java-API
And as before, /java-api/scalaris -w "rt" "4" -v gives
ConnectionFactory properties:
config file =
scalaris.node = [node1@domU-12-31-39-02-7E-DC.compute-1.internal]
scalaris.cookie = chocolate chip cookie
scalaris.client.name = java_client
scalaris.client.appendUUID = true
write failed with unknown:
de.zib.scalaris.UnknownException: Erlang message:
{badrpc,{'EXIT',{badarg,[{comm,send_local,2},{rdht_tx_read,work_phase,3},{rdht_t
x_write,work_phase,3},{rdht_tx,'-initiate_rdht_ops/1-lc$^0/1-0-',1},{rdht_tx,req
_list,2},{rdht_tx,req_list,2},{api_tx,req_list,2},{api_tx,write,2}]}}}
at de.zib.scalaris.CommonErlangObjects.processResult_commit(CommonErlangObjects.java:159)
at de.zib.scalaris.TransactionSingleOp.write(TransactionSingleOp.java:369)
at de.zib.scalaris.TransactionSingleOp.write(TransactionSingleOp.java:399)
at de.zib.scalaris.Main.main(Main.java:138)
Original comment by subhojit...@gmail.com
on 15 Dec 2011 at 9:55
oh, sorry, please start with
scalarisctl -f -s restart
unfortunately, scalarisctl is not that flexible regarding the order of its
commands (it specifically requires the format and order mentioned by
"scalarisctl -h").
btw: if you run into any problems, you can always check in our web debug
interface whether something went wrong, e.g. in this case,
http://localhost:8000/ring.yaws shows an empty ring.
Original comment by nico.kru...@googlemail.com
on 15 Dec 2011 at 10:02
FYI: regarding "scalarisctl checkinstallation" complaining about missing
Java-API files, this has been fixed in svn trunk - after installing all
dependencies (erlang, jdk, ant) try this:
svn checkout http://scalaris.googlecode.com/svn/trunk/ scalaris
cd scalaris
./configure --prefix=$HOME/temp
make && make java
make install && make install-java
cd ..
$HOME/temp/bin/scalarisctl checkinstallation
Original comment by nico.kru...@googlemail.com
on 15 Dec 2011 at 11:43
When Starting scalaris using "sudo scalarisctl -f -s restart " I am getting
following:
(node1@master)1> [warn] [ CC ] Cannot accept msg for unknown named process
mgmt_server: {connect}
[warn] [ CC ] Cannot locally send msg to unknown named process mgmt_server:
{send_to_group_member,service_per_vm,{hi}}
[warn] dht_node Connection failed, drop message
{send_to_group_member,service_per_vm,{hi}}
[warn] "to > 127.0.0.1:14196" Connection failed, drop message
{send_to_group_member,service_per_vm,{hi}}
[warn] "to > 127.0.0.1:14197" Connection failed, drop message
{send_to_group_member,service_per_vm,{hi}}
[warn] "to > 127.0.0.1:14198" Connection failed, drop message
{send_to_group_member,service_per_vm,{hi}}
[warn] [ CC ] Cannot locally send msg to unknown named process mgmt_server:
{register,{{10,248,129,38},14195,<0.116.0>}}
[warn] dht_node_reregister Connection failed, drop message
{register,{{10,248,129,38},14195,<0.116.0>}}
My scalaris.local.cfg is :
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Settings for distributed Erlang
% (see cs_send.erl to switch)
% {boot_host, {boot,'boot@foo.bar.com'}}.
% {log_host, {boot_logger, 'boot@foo.bar.com'}}.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Settings for TCP mode.
% (see cs_send.erl to switch)
% Insert the appropriate IP-addresses for your setup
% as comma separated integers:
% IP Address, Port, and label of the boot server
{boot_host, {{10,248,129,38},14195,node1}}.
% IP Address, Port, and label of the log server
{log_host, {{127,0,0,1},14195,boot_logger}}.
and scalaris.cfg is:
{mgmt_server, {{10,248,129,38},14195,mgmt_server}}.
{known_hosts, [{{10,248,129,38},14195, service_per_vm}, {{127,0,0,1},14196,
service_per_vm},
{{127,0,0,1},14197, service_per_vm}, {{127,0,0,1},14198, service_per_vm}]}.
Original comment by subhojit...@gmail.com
on 20 Dec 2011 at 1:20
First of all: your scalaris should be running fine - those warnings are
harmless. Check with the web debug interface at http://127.0.0.1:8000
Note: you are not supposed to change scalaris.cfg. Make your changes to
scalaris.local.cfg and you will be fine as those values override the other ones.
Your scalaris.local.cfg seems a bit outdated though, e.g. it is "mgmt_server"
instead of "boot_host" now. Just place the two lines from your scalaris.cfg
into scalaris.local.cfg.
This aside, if you do not start a management server (-m), then you should set
mgmt_server to e.g. 'none' which will eliminate the first two and the last
warning.
The other warnings come from the fact that you provided a list of known_hosts
but three of them were not available at the time of the node startup. Since you
are starting the first node, simply ignore them.
You should have at least one known host of mgmt server available.
btw: this is an issue tracker, your last message and future ones are probably
better off in our mailing list at http://groups.google.com/group/scalaris
Original comment by nico.kru...@googlemail.com
on 20 Dec 2011 at 5:49
Original issue reported on code.google.com by
subhojit...@gmail.com
on 12 Dec 2011 at 5:59