Lix Li comments

Results 19 comments of


                                            Lix Li

2018，人生是一条蛇，它咬着自己的尾巴

你去凤巢了吗？

建议中文注释/文档改用utf8编码

同上，类linux下无法使用，🤣

Using mmap instead of allocating large buffer in memory

when series cardinality is too high, compaction will result in out-of-memory and terminate `influxd` progress.

merge history db file into a common db file to avoid db file booming

> I don't believe this is safe, nothing stops something else with that pid coming along while this is running. In addition there's nothing stopping this being run concurrently. 1....

merge history db file into a common db file to avoid db file booming

#443 this issue has the same problem with me. I used multi process mode a few days and I found the endpoint for scrape latency was becoming more and more...

merge history db file into a common db file to avoid db file booming

> We can't make any assumptions about how this code is used, and given it's in a multi-process environment it must avoid any race conditions. > > How would this...

merge history db file into a common db file to avoid db file booming

We should point out that this code should only be used in the multiprocess mode

merge history db file into a common db file to avoid db file booming

I have added process lock code which is from django.core.files.locks

merge history db file into a common db file to avoid db file booming

> What happens if a process exists, sets a gauge, exits, and a new process comes along with the same pid? ```python for s in metric.samples: name, labels, value, timestamp,...

merge history db file into a common db file to avoid db file booming

> > Process exit and get the same pid is a low probability event. > > It's a certainty though at scale. Yes，Gauge is really a problem in multi process...