Dieter Plaetinck
Dieter Plaetinck
any thoughts on how the performance of this compares for single channel, 100% of msg going into diskqueue vs into WAL, and 100% of reads going out of diskqueue vs...
@mreiferson is ack-on-sync something you think should make it into nsqd as well? i think it would be great if my producers could know a message is safely stored in...
@mreiferson I doesn't look like it.. or there is a misunderstanding. `(d *diskQueue) Put(data []byte)` writes to d.writeChan and returns when it reads from d.writeResponseChan. and those ops only have...
@mreiferson: does this mean that 1) if this is enabled, we have guaranteed FIFO message ordering semantics? 2) will consumers be able to seek to arbitrary positions? for example if...
1. oh yes sure, i just meant single nsqd and assuming no requeues. a bit of a bold assumption i know.. let's call it "best case scenario" ;-) 2. cool...
@mreiferson new years resolution to pick this back up ? :)
> I suggest a change to bra such that if any inotify commands come in while the command is running, the command is run again after it finishes. After some...
When I reorganize the code a bit, it works fine: ``` $ pwd /home/dieter/go/src/github.com/Dieterbe/deadtest $ find . . ./main.go ./foo ./foo/f.go ./foo/f_test.go $ cat main.go package main import ( "fmt"...
i guess that would be nice but i don't have time to work on that.
this code change would hide problems where your first outage event is not a start event. instead of silently allowing this case, we should actually even react stronger and report...