Wizard_
Members
-
Joined
-
Last visited
Solutions
-
Wizard_'s post in Recurring HDD spin‑ups after adding ZFS SSD was marked as the answerAfter several days of observation, it can be confirmed that the issue is still related to atime. To avoid disk wake-ups caused by atime, all datasets must have atime manually set to off. Inheriting the pool’s setting will still cause wake-ups, even if the pool itself is configured with atime=off.
-
Wizard_'s post in 怎么升级docker engine？ was marked as the answerunraid7已经把docker版本升级到25以上了，如果实在有需要可以尝试刚出的unraid7
-
Wizard_'s post in 直通显卡后cpu长期满载占用，unraid未死机但无法进行操作 was marked as the answer似乎问题解决了？具体我也不太清楚是什么情况，但是貌似"无法纠正的I/O错误"问题不再出现、RxErr问题和BadDLLP问题无论高低负载情况都没有再出现。
似乎之前日志内要求的两条启动参数解决了问题。
此问题暂时解决，如有其他问题我会更新。
简单总结一下解决方案：
在syslinux内加上两条启动参数，即
nvme_core.default_ps_max_latency_us=0 pcie_aspm=off
命令作用：
第一条命令的作用是禁止nvme的APST，禁止ssd主动进入省电模式，只能进入Host设置的电源状态。
第二条命令的作用是禁止linux内核的ASPM，在这种情况下系统将不会启用ASPM，也就不会主动省电。
-
Wizard_'s post in 几个更换主板＆升级系统后出现的问题 was marked as the answer此问题通过重装中文语言包解决
-
Wizard_'s post in 关于unraid的网络问题 was marked as the answer问题解决，已经数天多次重启未复现问题。
应该是passwall代理出了问题，通过修改TCP转发端口端口实现
22,25,53,143,465,587,853,993,995,80,443 此为应该代理的端口