Re: MySQL Deadlocks in OfBiz 12.04

2017-02-18 Thread Scott Gray
I don't know, I'm inclined to think these lists do a better job of sharing information in a searchable manner than Jira does. Regards Scott On 19/02/2017 00:36, "Jacques Le Roux" wrote: Thanks Scott, As I referred to the thread it could help people to fix related

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-18 Thread Jacques Le Roux
Thanks Scott, As I referred to the thread it could help people to fix related issues if they stumble upon. Else if you think it's confusing, please feel free to change the summary and description, maybe by adding your comment below? Jacques Le 18/02/2017 à 11:37, Scott Gray a écrit : There

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-18 Thread Scott Gray
There was no bug report in this thread, just some vague details. Deadlocks are not simple to locate and fix, finding the competing transactions is difficult even with access to the logs and database. I don't think there's any point in filing a ticket unless Robb comes back with more details

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-18 Thread Jacques Le Roux
Hi Scott, I was referring to the bug which was initially reported in this thread. I understand that we should rather refactor FinAccount services and I have opened a Jira for that https://issues.apache.org/jira/browse/OFBIZ-9221 Jacques Le 18/02/2017 à 08:06, Scott Gray a écrit : Which

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-17 Thread Scott Gray
Which bug are you referring to Jacques? FinAccount services certainly need to be refactored but the use case that I mentioned was unusual so it probably isn't worth fixing unless a more common scenario comes along. It just happened to be a simple example I had on hand. Regards Scott On

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-17 Thread Pierre Smits
+1. Feel free to investigate and create when applicable. Best regards, Pierre On Friday, February 17, 2017, Jacques Le Roux wrote: > I mean if it still exists in trunk and R16.11 > > > Le 17/02/2017 à 13:51, Jacques Le Roux a écrit : > >> Hi Scott, >> >> Should

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-17 Thread Jacques Le Roux
I mean if it still exists in trunk and R16.11 Le 17/02/2017 à 13:51, Jacques Le Roux a écrit : Hi Scott, Should we not open a Jira and try to fix this bug? Jacques Le 16/02/2017 à 04:04, Scott Gray a écrit : For example, a current project has some heavily used FinAccount rows and the

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-17 Thread Jacques Le Roux
Hi Scott, Should we not open a Jira and try to fix this bug? Jacques Le 16/02/2017 à 04:04, Scott Gray a écrit : For example, a current project has some heavily used FinAccount rows and the OFBiz implementation has a tendency to modify the child rows (FinAccountAuth/FinAccountTrans) before

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-17 Thread Robb Wagoner
Thanks Scott! Much appreciated. The pattern makes sense. Have a great weekend. On Wed, Feb 15, 2017 at 8:05 PM Scott Gray wrote: For example, a current project has some heavily used FinAccount rows and the OFBiz implementation has a tendency to modify the child

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-16 Thread Jacques Le Roux
Then you have not the infamous problem of ms https://issues.apache.org/jira/browse/OFBIZ-2208 and it's another issue indeed Jacques Le 16/02/2017 à 00:08, Robb Wagoner a écrit : It is AWS Aurora variant of MySQL v5.6.10. $ mysql -hofbiz.example.com -e 'select version(), @@aurora_version'

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-15 Thread Scott Gray
For example, a current project has some heavily used FinAccount rows and the OFBiz implementation has a tendency to modify the child rows (FinAccountAuth/FinAccountTrans) before modifying the parent FinAccount row. This very easily results in a deadlock as soon as one of the competing

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-15 Thread Shi Jinghai
:). -邮件原件- 发件人: Robb Wagoner [mailto:r...@pandastrike.com] 发送时间: 2017年2月16日 7:09 收件人: user@ofbiz.apache.org 主题: Re: MySQL Deadlocks in OfBiz 12.04 It is AWS Aurora variant of MySQL v5.6.10. $ mysql -hofbiz.example.com -e 'select version(), @@aurora_version

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-15 Thread Robb Wagoner
It is AWS Aurora variant of MySQL v5.6.10. $ mysql -hofbiz.example.com -e 'select version(), @@aurora_version' ++--+ | version() | @@aurora_version | ++--+ | 5.6.10-log | 1.10 | ++--+ On Wed, Feb 15,

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-15 Thread Jacques Le Roux
Hi Robb, Just to be sure, the MySQL version you use >= 5.6.4 ? Jacques Le 15/02/2017 à 20:03, Robb Wagoner a écrit : We are running OfBiz 12.04 and are using MySQL 5.6 as the RDBMS. We are experiencing database deadlocks and would like to be able to add MySQL-specific statements to deal with

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-15 Thread Robb Wagoner
Thanks Scott. Since I am new to OfBiz can you provide a code example? I.e. by what means are you issuing an arbitrary no-op update statement on a parent row? On Wed, Feb 15, 2017 at 1:02 PM Scott Gray wrote: Hi Robb, I've encountered similar deadlocks in the past

Re: MySQL Deadlocks in OfBiz 12.04

2017-02-15 Thread Scott Gray
Hi Robb, I've encountered similar deadlocks in the past and have simply issued an update statement on the main row before attempting to work with any child rows. That has the same effect as the FOR UPDATE statement but without needing to enhance the framework. Regards Scott On 16/02/2017 08:03,

MySQL Deadlocks in OfBiz 12.04

2017-02-15 Thread Robb Wagoner
We are running OfBiz 12.04 and are using MySQL 5.6 as the RDBMS. We are experiencing database deadlocks and would like to be able to add MySQL-specific statements to deal with the deadlocks. For example: `SELECT ... FOR UPDATE` on PRIMARY KEYs and/or FOREIGN KEYs in the transactions. Being new to