[pgpool-hackers: 1086] Re: Bugs/issues pending

[Tatsuo Ishii]

Ahsan,

Thank you for checking the list. Here are my commnets.

> On Wed, Oct 7, 2015 at 7:00 AM, Tatsuo Ishii <ishii at postgresql.org> wrote:
> 
>> Here is the list of bugs/issunes needed to be addressed before 3.5 is
>> out. The numbers in the first column in each row is for my management
>> purpose. "bug#xxx" denotes the id in our bug tracker.
>>
> 
> Thanks for sharing the list. Please see my feedback inline...we need to
> come up with list of issues that we feel are must-have for 3.5 beta. I
> don't think we should work on any new features, just try to complete that
> feature that already in progress...
> 
> 
>> Let me know if there's anything I missed.
>>
>>         ID      Date    Person in charge        Overview        Status
>>
>> 5       bug#135 2015/05/20      Usama   delegate ip does not up assigned
>>
>>         The last feed back from the reporter is as follows. Since the
>>         installer is maintained by Nozomi, I think it's better to
>>         change the person in charge to Nozomi.
>>
>>
> Makes sense..
> 
> 
>>         2015-05-21 21:40
>>         I believe this problem is an installer
>>         bug. installer2-pg92-3.4.0 in ./lib/pgpool.sh it does: for
>>         host 0 _writePgpoolParam heartbeat_destination0
>>         "'${PGPOOL_HOST_ARR[0]}'" for host 1 _writePgpoolParam
>>         heartbeat_destination0 "'${PGPOOL_HOST_ARR[0]}'"
>>
>> 6       bug#138 2015/07/19      Tatsuo  address already used    feedback
>>
>>         I got no resoponse from the original reporter since Sep 17,
>>         2015 so far. I think we can close the item.
>>
> 
> okay...
> 
>>
>> 7       bug#139 2015/07/17      Usama   broken arping_cmd       assigned
>>
>>         Need to care.
>>
> 
> I haven't looked closely...is that also watchdog?

Yes, I thinkk so.

------------------------------------------------------------------------
0000139: broken arping_cmd doesn't get reported in pgpool 3.3.6
Description	When having a misconfigured arping_cmd this isn't caught by pgpool 3.3.6:

Jul 17 12:16:42 off-zk-02-ber pgpool: 2015-07-17 12:16:42 DEBUG: pid 21801: exec_ifconfig: 'pgpool_ifconfig eth0:pgpool2-slow inet $_IP_$ netmask 255.255.255.0' succeeded
Jul 17 12:16:42 off-zk-02-ber pgpool: 2015-07-17 12:16:42 DEBUG: pid 21801: exec_ifconfig: 'arping -U -I eth0 -c 10 $_IP_$' succeeded
[..]
Jul 17 12:16:44 off-zk-02-ber pgpool: 2015-07-17 12:16:44 LOG: pid 21801: wd_IP_up: ifconfig up succeeded

But when looking at a strace of the watchdog process it clearly shows the misconfigured arping path:

execve("/usr/sbin/arping", ["arping", "-U", "-I", "eth0", "-c", "10", "192.168.181.232"], [/* 18 vars */]) = -1 ENOENT (No such file or directory)
------------------------------------------------------------------------

>> 10      bug#141 2015/08/13      Usama   watchdog        assigned
>>
>>         Need to care.
>>
> 
> Yes this should be fixed with 3.5 beta
> 
>>
>> 11      pgpool-general:3908     2015/08/02      Yugo    unexpected
>> rollback     feedback
>>
>>         No response from the original reporter since Aug 4, 2015. I
>>         think we can close the item.
>>
>> 12      pgpool-general:3907     2015/08/02      Usama   watchdog info
>>  assigned
>>
>>         Need to care.
>>
>> Yes this should be fixed with 3.5 beta
> 
> 
>> 14      pgpool-general:3923     2015/08/06      Tatsuo  pgpool_status
>>  needs discussion
>>
>>         Not bug but feature. Will start discussion soon.
>>
>>
> This is not must-have for 3.5 beta...

Let me think about this...

>> 18      pgpool-hackers:994      2015/08/05      Tatsuo  connect to
>> necessary DB nodes   proposal, feedback
>>
>>         Waiting for feedback.
>>
>> This is not must-have for 3.5 beta...

Ok, leave it for 3.6.

>> 21      pgpool-general:4006     2015/08/27      Nozomi  pgpoolAdmin
>>  assigned
>>
>>         Feedback from the original reporter on Sep 05, 2015. Waiting
>>         for Nozomi's response.
>>
>> 23      pgpool-hackers:1034     2015/09/03      -       add new child if
>> it goes down   need to assign
>>
>>
> This is not must-have for 3.5 beta?

No, but current behavior of pgpool is painful for users.

>> 24      pgpool-general:4016     2015/09/02      -       deal with
>> pg_terminate_backend  need to assign or expecting patches from user
>>
>>
> This is not must-have for 3.5 beta...

Yeah, we haven't had any patches from user...

>> 25      pgpool-hackers:1035     2015/09/03      Yugo    assigned
>>
>>         > Currently build farm tests include following combinations:
>>         > Branches: master, 3.4-stable, 3.3-stable
>>         > PostgreSQL: 9.3.9
>>         > Distributions: CentOS 6.7
>>         >
>>         > Probably we should add more combinations. For example:
>>         > PostgreSQL 9.4
>>         > CentOS 7
>>
>>         Even we should have PostgreSQL 9.5 since pgpool-II 3.5 has
>>         support for PostgreSQL 9.5 parser.
>>
>>
> Yes it is nice to have for 3.5 beta..

Best regards,
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese:http://www.sraoss.co.jp