是否有可能加速“死亡”的消除。来自假脱机表的消息？ [英] Is it possible to speed up the removal of "dead" messages from the spool table?

问题描述

我有一个解决方案，每天处理数千条消息（很快就会达到1米以上）。经过大量的实验和调试后，我对吞吐量感到满意，但我最近注意到假脱机表是滞后的，现在看起来它删除了死信息比新信息更慢。显然这令人担忧 - MsgBox一度达到了72GB！

SQL清理工作似乎每秒消耗大约20-40个死信息，但我们可能持续超过50 msg /秒，因此它永远不会赶上。

I have a solution which processes many thousands of messages a day (soon to be 1m+). After much experimentaion and tuning I'm happy with the throughput, but I recently noticed the spool table is lagging, and now looks like it removes dead message slower than new ones arrive. Clearly this is worrying - the MsgBox at one point reached 72GB !!

The SQL cleanup jobs seem to polish of about 20-40 or so dead messages a second, but we probably have a sustained rate of over 50 msg/second, so it never catches up.

I不要相信硬件正在挣扎（虽然我不确定在SQL中查看哪些perf计数器以了解为什么它不能更快​​地处理），所以任何线索？

I don't believe the hardware is struggling (although I'm not sure what perf counters to look at in SQL to see why it can't process faster), so any clues?

问候，Jason

Regards, Jason

推荐答案

您已查看并配置了SQL代理中安装的所有SQL作业？

You have looked and configured all of the SQL Jobs that are installed in the SQL Agent?

这篇关于是否有可能加速“死亡”的消除。来自假脱机表的消息？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！