|
本帖最后由 金鹏 于 2011-12-29 15:27 编辑
Thu Dec 29, 2011 1:12 am
update on bigadv
We wanted to give a quick update on bigadv. No new policies or changes announced here. The purpose of this post is mostly to be a little more transparent about what we've been up to (and why people may have seen changes in bigadv availability).
Briefly, one of the servers ran into a code bug ("feature"), and the fix I hacked for it caused other problems, so one of the bigadv servers has been largely offline. That server was handling a lot of the bigadv-8 traffic, resulting in more limited availability of bigadv-8 work units.
I have personally been busy with other server transitions, particularly moving data and virtual "servers" from Stanford to new machines I have acquired at Virginia. I have a new bigadv server lined up, but we've also had some issues with the physical installation there (the RAID arrays shipped with sub-standard rails, which is annoying). We're hoping to resolve those early in January, and once that server is up and running, I'll start prototyping projects for bigadv-16. More broadly, we're increasing the geographical distribution of FAH servers, which should help a lot with redundancy at the times (fortunately infrequent) when we have large-scale outages at Stanford.
The bottom line is that we haven't been intentionally sunsetting the bigadv-8 projects, but we've had a confluence of bigadv-8 supply and server code issues at the same time that we've been busy with other server transitions. The policy plan remains to bring new bigadv-16 projects online and then sunset bigadv-8 no sooner than Jan 15.
大意:
bigadv消息更新
有一台服务器代码有错,修复bug时又产生了新的问题,导致一台bigadv服务器宕机。这台机子主要负责8核任务,所以现在8核任务很少了。
现在主要工作是服务器迁移,把服务器和数据从Stanford迁移到Virginia。一台新BA服务器已经到位,不过硬件安装出了问题(RAID系统安装出错)。我们希望能在1月解决这个问题。那台服务器一但装好,我们就会开始测试BA-16任务。我们准备把服务器尽量分散开(不把鸡蛋放在一个篮子里),以避免Stanford网络出问题的时候,服务器总是被‘一锅端’。
再次强调,我们不是故意把BA-8下线的,实属意外。本来BA-8的任务就少,再加上服务器代码出错,忙于服务器迁移工作,真的是屋漏逢夜雨啊。不过我们还是要说明,BA-16的上线日期以及BA-8的下线日期,不会早于1月15日。
vmzy 发表于 2011-12-29 11:14 |
|