You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@servicecomb.apache.org by Willem Jiang <wi...@gmail.com> on 2020/05/05 12:30:34 UTC
Re: [PACK][DISCUSS] Resume suspended global transaction
Hi ZhangLei,
Is there any update of this issue?
Willem Jiang
Twitter: willemjiang
Weibo: 姜宁willem
On Sat, Apr 18, 2020 at 3:44 PM Willem Jiang <wi...@gmail.com> wrote:
>
> For the case 1， I think the better way is keep the transaction failed there.
> For the case 2, we could retry the compensation call if the
> compensation method is idempotent.
> For the case 3, there are something wrong with the processing, I think
> it's better to suspend the transaction processing.
>
> Willem Jiang
>
> Twitter: willemjiang
> Weibo: 姜宁willem
>
> On Sat, Apr 18, 2020 at 12:18 AM Zhang Lei <co...@gmail.com> wrote:
> >
> > Hi Team,
> >
> > Current, the following case will cause global transactions to suspend
> >
> > case 1. Global transaction timeout
> > case 2. Compensate failed
> > case 3. Unpredictable sequence of events
> >
> > Solution
> >
> > case 1: When subsequent events are received, it can resume to the state
> > before suspension and continue to execute
> > case 2: Playback the last event of the suspend global transactions by
> > triggering in the Alpha UI, it can resume to the state before suspension
> > and continue to compensate
> > case 3: Do you have any suggestions? Or Still keep suspended.
> >
> >
> > Best regards,
> > Lei Zhang
Re: [PACK][DISCUSS] Resume suspended global transaction
Posted by Zhang Lei <co...@gmail.com>.
Hi Willem,
I am trying to solve the case 2, before a little busy recently, this might
take some time.
Best regards,
Lei Zhang
On May 5, 2020 at 8:30:50 PM, Willem Jiang (willem.jiang@gmail.com) wrote:
Hi ZhangLei,
Is there any update of this issue?
Willem Jiang
Twitter: willemjiang
Weibo: 姜宁willem
On Sat, Apr 18, 2020 at 3:44 PM Willem Jiang <wi...@gmail.com>
wrote:
>
> For the case 1， I think the better way is keep the transaction failed
there.
> For the case 2, we could retry the compensation call if the
> compensation method is idempotent.
> For the case 3, there are something wrong with the processing, I think
> it's better to suspend the transaction processing.
>
> Willem Jiang
>
> Twitter: willemjiang
> Weibo: 姜宁willem
>
> On Sat, Apr 18, 2020 at 12:18 AM Zhang Lei <co...@gmail.com> wrote:
> >
> > Hi Team,
> >
> > Current, the following case will cause global transactions to suspend
> >
> > case 1. Global transaction timeout
> > case 2. Compensate failed
> > case 3. Unpredictable sequence of events
> >
> > Solution
> >
> > case 1: When subsequent events are received, it can resume to the state
> > before suspension and continue to execute
> > case 2: Playback the last event of the suspend global transactions by
> > triggering in the Alpha UI, it can resume to the state before
suspension
> > and continue to compensate
> > case 3: Do you have any suggestions? Or Still keep suspended.
> >
> >
> > Best regards,
> > Lei Zhang