[Postgres-xl-developers] fix postgres-xl build on FreeBSD

Jov amutu at amutu.com
Thu Dec 11 00:01:22 PST 2014


I will put the patch to pgxl,but the review progress is slow.there is no
feedback from Mason Sharp for the first patch by now.
I will maintain the patch and try to create a FreeBSD ports for pgxl recent.

We are heavily test the pgxl for more than 4 months and find some bugs,some
of the bugs fix patch have already put to the community by my college
jasonysli.We fix bug quickly in our team,but the community review progress
is slow.How can I help to improve this situation?

Jov
blog: http:amutu.com/blog <http://amutu.com/blog>

2014-12-11 15:28 GMT+08:00 Ramanuj Hari <hari.ramanuj at gmail.com>:

> Hi Jov,
>
> Thanks a lot for patch.
>
> I have applied the patch on "Postgres-XL 9.2rc" and it works fine on
> FreeBSD 8.4 System.
>
> Have tested with 1 GTM, 2 Co-ordinator, 1 GTM Proxy, 4 Datanode
> configuration.
>
> Our company has plan to provide scalable database solution based on
> Postgres-XL for Analytics product to be released in Q2-2015.
>
> Hi,
>
> It would be great if you can review patch and provide comments if any.
>
> Thanks,
> Ramanuj
>
>
>
> On Wed, Dec 10, 2014 at 9:46 PM, Ramanuj Kumar <hari.ramanuj at gmail.com>
> wrote:
>
>> Hi Jov,
>>
>> Thanks a lot for the patch. Let me try that out.
>>
>> Yes, coord log has error message about signal 10 and “bus error”.
>>
>> —full coord log—
>> 2014-12-09 22:52:34 PST [] [1175]: [1-1] user=,db=,e=00000 LOG:  database
>> system was shut down at 2014-12-09 22:52:31 PST
>> 2014-12-09 22:52:34 PST [] [1170]: [1-1] user=,db=,e=00000 LOG:  database
>> system is ready to accept connections
>> 2014-12-09 22:52:34 PST [] [1180]: [1-1] user=,db=,e=00000 LOG:
>>  autovacuum launcher started
>> 2014-12-09 22:52:44 PST [10.102.103.130(19381)] [1348]: [1-1]
>> user=pgxl,db=postgres,e=00000 LOG:  statement: ALTER NODE coord1 WITH
>> (HOST='10.102.103.130', PORT=11921);
>> 2014-12-09 22:52:44 PST [10.102.103.130(19381)] [1348]: [2-1]
>> user=pgxl,db=postgres,e=00000 LOG:  duration: 1.465 ms
>> 2014-12-09 22:52:44 PST [10.102.103.130(19381)] [1348]: [3-1]
>> user=pgxl,db=postgres,e=00000 LOG:  statement: CREATE NODE datanode1 WITH
>> (TYPE='datanode', HOST='10.102.103.130', PORT=12921, PRIMARY, PREFERRED);
>> 2014-12-09 22:52:44 PST [10.102.103.130(19381)] [1348]: [4-1]
>> user=pgxl,db=postgres,e=00000 LOG:  duration: 1.388 ms
>> 2014-12-09 22:52:44 PST [10.102.103.130(16075)] [1352]: [1-1]
>> user=pgxl,db=postgres,e=00000 LOG:  statement: EXECUTE DIRECT ON
>> (datanode1) 'CREATE NODE coord1 WITH (TYPE=''coordinator'',
>> HOST=''10.102.103.130'', PORT=1
>> 1921)';
>> 2014-12-09 22:52:44 PST [10.102.103.130(16075)] [1352]: [2-1]
>> user=pgxl,db=postgres,e=00000 LOG:  Connection error could not send data to
>> server
>> 2014-12-09 22:52:44 PST [10.102.103.130(16075)] [1352]: [3-1]
>> user=pgxl,db=postgres,e=00000 STATEMENT:  EXECUTE DIRECT ON (datanode1)
>> 'CREATE NODE coord1 WITH (TYPE=''coordinator'', HOST=''10.102.103.130'',
>> PORT=11921)
>> ';
>> 2014-12-09 22:52:44 PST [10.102.103.130(16075)] [1352]: [4-1]
>> user=pgxl,db=postgres,e=00000 LOG:  Connection error failed to send data to
>> datanode
>> 2014-12-09 22:52:44 PST [10.102.103.130(16075)] [1352]: [5-1]
>> user=pgxl,db=postgres,e=00000 STATEMENT:  EXECUTE DIRECT ON (datanode1)
>> 'CREATE NODE coord1 WITH (TYPE=''coordinator'', HOST=''10.102.103.130'',
>> PORT=11921)
>> ';
>> 2014-12-09 22:52:44 PST [] [1170]: [2-1] *user=,db=,e=00000 LOG:  server
>> process (PID 1352) was terminated by signal 10: Bus error*
>> 2014-12-09 22:52:44 PST [] [1170]: [3-1] user=,db=,e=00000 DETAIL:
>>  Failed process was running: EXECUTE DIRECT ON (datanode1) 'CREATE NODE
>> coord1 WITH (TYPE=''coordinator'', HOST=''10.102.103.130'', PORT=11921)';
>> 2014-12-09 22:52:44 PST [] [1170]: [4-1] user=,db=,e=00000 LOG:
>>  terminating any other active server processes
>> 2014-12-09 22:52:44 PST [] [1180]: [2-1] user=,db=,e=57P02 WARNING:
>>  terminating connection because of crash of another server process
>> 2014-12-09 22:52:44 PST [] [1180]: [3-1] user=,db=,e=57P02 DETAIL:  The
>> postmaster has commanded this server process to roll back the current
>> transaction and exit, because another server process exited abnormally and p
>> ossibly corrupted shared memory.
>> 2014-12-09 22:52:44 PST [] [1180]: [4-1] user=,db=,e=57P02 HINT:  In a
>> moment you should be able to reconnect to the database and repeat your
>> command.
>> 2014-12-09 22:52:44 PST [] [1170]: [5-1] user=,db=,e=00000 LOG:  all
>> server processes terminated; reinitializing
>> 2014-12-09 22:52:44 PST [] [1357]: [1-1] user=,db=,e=00000 LOG:  database
>> system was interrupted; last known up at 2014-12-09 22:52:34 PST
>> 2014-12-09 22:52:44 PST [] [1357]: [2-1] user=,db=,e=00000 LOG:  database
>> system was not properly shut down; automatic recovery in progress
>> 2014-12-09 22:52:44 PST [] [1357]: [3-1] user=,db=,e=00000 LOG:  redo
>> starts at 0/186B110
>> 2014-12-09 22:52:44 PST [] [1357]: [4-1] user=,db=,e=00000 LOG:  record
>> with zero length at 0/186B660
>> 2014-12-09 22:52:44 PST [] [1357]: [5-1] user=,db=,e=00000 LOG:  redo
>> done at 0/186B630
>> 2014-12-09 22:52:44 PST [] [1357]: [6-1] user=,db=,e=00000 LOG:  last
>> completed transaction was at log time 2014-12-09 22:52:44.820342-08
>> 2014-12-09 22:52:44 PST [] [1361]: [1-1] user=,db=,e=00000 LOG:
>>  autovacuum launcher started
>> 2014-12-09 22:52:44 PST [] [1170]: [6-1] user=,db=,e=00000 LOG:  database
>> system is ready to accept connections
>> —————
>>
>> Thanks,
>> Ramanuj
>>
>>
>>
>>
>> On 10-Dec-2014, at 9:41 pm, Jov <amutu at amutu.com> wrote:
>>
>> the whole patch:
>> http://amutu.com/blog/public_stage/xl_freebsd.patch
>>
>> this patch include the above patch,and also fix the get/send invalid fds
>> from pooler process,and some new dtrace probe.
>>
>> you can check the coord logs,and find if there are some error message
>> about "signal 11",or "bus error"
>>
>>
>> Jov
>> blog: http:amutu.com/blog <http://amutu.com/blog>
>>
>> 2014-12-10 19:53 GMT+08:00 Ramanuj Hari <hari.ramanuj at gmail.com>:
>>
>>> Hi Jov,
>>>
>>> It would be great, if you could please share your other patch with me. I
>>> will try if that works on my environment.
>>>
>>> Thanks,
>>> Ramanuj
>>>
>>>
>>> On Wed, Dec 10, 2014 at 3:52 PM, Jov <amutu at amutu.com> wrote:
>>>
>>>> have you test with the patch above?I install pgxl with pgxc_ctl.
>>>> I run PG-XL on FreeBSD with some more patch.The patch above can only
>>>> make pgxc_ctl install and init cluster,but when you run SQL,there will be
>>>> problem.I have solve the problem,and wait for the above patch accepting and
>>>> then will submit the other patch.
>>>>
>>>> for install and init pgxl,there are 2 problems for FreeBSD user to
>>>> solve:
>>>> 1.install the bash shell and apply the above patch.
>>>> 2.set remote env for ssh.you should add a file named environment in
>>>> the ~/.ssh dir with PATH set to pgxl bin dir.
>>>>
>>>> Jov
>>>> blog: http:amutu.com/blog <http://amutu.com/blog>
>>>>
>>>> 2014-12-10 17:56 GMT+08:00 Ramanuj Hari <hari.ramanuj at gmail.com>:
>>>>
>>>>> Hi Jov,
>>>>>
>>>>> I am also trying to run on FreeBSD r10. But got into issue.
>>>>>
>>>>> ----
>>>>> do_shell.c:dump_cmdList(713) *** cmdList Dump
>>>>> *******************************
>>>>> allocated = 2, used = 1
>>>>> do_shell.c:dump_cmdList(723) === CMD: 0 ===
>>>>> do_shell.c:dump_cmdList(727)    --- CMD-EL: 0:host="NULL",
>>>>> command="psql -p 11921 -h 10.102.103.135 -a postgres pgxl",
>>>>> localStdin="/tmp/STDIN_48989_34", localStdout="NULL"
>>>>> EXECUTE DIRECT ON (datanode1) 'CREATE NODE coord1 WITH
>>>>> (TYPE=''coordinator'', HOST=''10.102.103.135'', PORT=11921)';
>>>>> EXECUTE DIRECT ON (datanode1) 'ALTER NODE datanode1 WITH
>>>>> (TYPE=''datanode'', HOST=''10.102.103.135'', PORT=12921, PRIMARY,
>>>>> PREFERRED)';
>>>>> EXECUTE DIRECT ON (datanode1) 'SELECT pgxc_pool_reload()';
>>>>> do_shell.c:dump_cmdList(737)    ----------
>>>>> do_shell.c:doCmdEl(360) Local command: "psql -p 11921 -h
>>>>> 10.102.103.135 -a postgres pgxl", actual: "( psql -p 11921 -h
>>>>> 10.102.103.135 -a postgres pgxl ) > /tmp/STDOUT_48989_35 2>&1 <
>>>>> /tmp/STDIN_48989_34"
>>>>> EXECUTE DIRECT ON (datanode1) 'CREATE NODE coord1 WITH
>>>>> (TYPE=''coordinator'', HOST=''10.102.103.135'', PORT=11921)';
>>>>> ERROR:  Could not begin transaction on Datanodes
>>>>> EXECUTE DIRECT ON (datanode1) 'ALTER NODE datanode1 WITH
>>>>> (TYPE=''datanode'', HOST=''10.102.103.135'', PORT=12921, PRIMARY,
>>>>> PREFERRED)';
>>>>> ERROR:  Could not begin transaction on Datanodes
>>>>> EXECUTE DIRECT ON (datanode1) 'SELECT pgxc_pool_reload()';
>>>>> ERROR:  Could not begin transaction on data node.
>>>>> coord_cmd.c:configure_datanodes(416) Done.
>>>>> PGXC$
>>>>> -----
>>>>>
>>>>> Could you please help how did you try, whether it is pgxc_ctl or
>>>>> manual configuration.
>>>>>
>>>>> Thanks,
>>>>> Ramanuj
>>>>>
>>>>>
>>>>> On Wed, Oct 15, 2014 at 7:21 AM, Jov <amutu at amutu.com> wrote:
>>>>>
>>>>>> v2 change:
>>>>>> replace "#!/bin/bash" with "#!/usr/bin/env bash" in these files:
>>>>>> contrib/pgxc_ctl/pgxc_ctl.bash
>>>>>> contrib/pgxc_ctl/pgxc_ctl_bash.c
>>>>>> contrib/pgxc_ctl/pgxc_ctl_bash.org
>>>>>> contrib/pgxc_ctl/pgxc_ctl_bash_2
>>>>>> contrib/pgxc_ctl/pgxc_ctl_conf_part
>>>>>>
>>>>>> I am success to setup and init pgxl cluster after this patch on
>>>>>> FreeBSD r10 amd64
>>>>>>
>>>>>> Jov
>>>>>> blog: http:amutu.com/blog <http://amutu.com/blog>
>>>>>>
>>>>>> 2014-10-15 5:52 GMT+08:00 Jov <amutu at amutu.com>:
>>>>>>
>>>>>>> Thanks.I find some runtime problem when running pgxc_ctl,I will
>>>>>>> submit a updated patch later.
>>>>>>> 2014年10月15日 4:52 AM于 "Mason Sharp" <msharp at translattice.com>写道:
>>>>>>>
>>>>>>> Hi Jov,
>>>>>>>>
>>>>>>>> Will try to take a closer look soon.
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>>
>>>>>>>> Mason
>>>>>>>>
>>>>>>>> On Mon, Oct 13, 2014 at 2:50 AM, Jov <amutu at amutu.com> wrote:
>>>>>>>>
>>>>>>>>> any idea?
>>>>>>>>>
>>>>>>>>> Jov
>>>>>>>>> blog: http:amutu.com/blog <http://amutu.com/blog>
>>>>>>>>>
>>>>>>>>> 2014-10-11 17:41 GMT+08:00 赵伟 <amutu at amutu.com>:
>>>>>>>>>
>>>>>>>>>> when build pgxl on FreeBSD,there are 3 problems:
>>>>>>>>>> 1."WEXITSTATUS(status) " macro can't be resolve when compile
>>>>>>>>>> src/gtm/gtm_ctl/gtm_ctl.c ,caused by lacking of sys/wait.h header.
>>>>>>>>>>
>>>>>>>>>> 2.type "pid_t" undefined when compile
>>>>>>>>>> contrib/pgxc_ctl/utils.h,caused by lacking of sys/types.h header
>>>>>>>>>>
>>>>>>>>>> 3.contrib/pgxc_ctl/make_signature bash can't be running bacause
>>>>>>>>>> of it think the bash is installed at /bin/bash,which is not true on
>>>>>>>>>> FreeBSD.replace with "#!/usr/bin/env bash" solve the problem.
>>>>>>>>>>
>>>>>>>>>> patch attached.
>>>>>>>>>>
>>>>>>>>>> thanks,
>>>>>>>>>> Jov
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>> Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer
>>>>>>>>>> Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS
>>>>>>>>>> Reports
>>>>>>>>>> Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White
>>>>>>>>>> paper
>>>>>>>>>> Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog
>>>>>>>>>> Analyzer
>>>>>>>>>> http://p.sf.net/sfu/Zoho
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Postgres-xl-developers mailing list
>>>>>>>>>> Postgres-xl-developers at lists.sourceforge.net
>>>>>>>>>>
>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/postgres-xl-developers
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>> Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer
>>>>>>>>> Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS
>>>>>>>>> Reports
>>>>>>>>> Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White
>>>>>>>>> paper
>>>>>>>>> Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog
>>>>>>>>> Analyzer
>>>>>>>>> http://p.sf.net/sfu/Zoho
>>>>>>>>> _______________________________________________
>>>>>>>>> Postgres-xl-developers mailing list
>>>>>>>>> Postgres-xl-developers at lists.sourceforge.net
>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/postgres-xl-developers
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> Mason Sharp
>>>>>>>>
>>>>>>>> TransLattice - http://www.translattice.com
>>>>>>>> Clustered and Distributed Database Solutions
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Comprehensive Server Monitoring with Site24x7.
>>>>>> Monitor 10 servers for $9/Month.
>>>>>> Get alerted through email, SMS, voice calls or mobile push
>>>>>> notifications.
>>>>>> Take corrective actions from your mobile device.
>>>>>> http://p.sf.net/sfu/Zoho
>>>>>> _______________________________________________
>>>>>> Postgres-xl-developers mailing list
>>>>>> Postgres-xl-developers at lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/postgres-xl-developers
>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.postgres-xl.org/private.cgi/postgres-xl-developers-postgres-xl.org/attachments/20141211/5d44a167/attachment.htm>


More information about the Postgres-xl-developers mailing list