For a few weeks now, I have noticed my Gitea instance takes nearly two minutes to load.
Turns out, when you use gitea for archiving lots of public repos, this will build up many entries in the actions table, which can drastically increase the loading time needed.
To fix this, I made a query & cron job, which removes actions for my "Archive" organization.
This- fixes the issue.
IF, you aren't archiving/mirroring a lot of repositories, this post will not help you.
Copy/paste the below script into an administrative command prompt. Thats it.
:: Set "Old" Explorer Context Menu as Default
reg add "HKEY_CURRENT_USER\SOFTWARE\CLASSES\CLSID\{86ca1aa0-34aa-4e8b-a509-50c905bae2a2}\InprocServer32" /ve /f
:: Remove Explorer "Command Bar"
reg add "HKCU\Software\Classes\CLSID\{d93ed569-3b3e-4bff-8355-3c44f6a52bb5}\InprocServer32" /f /ve
:: Restart Windows Explorer. (Applies the above settings without needing a reboot)
taskkill /f /im explorer.exe
start explorer.exe
:: Empty Comment (Prevents you from having to press "enter" to execute the line to restart explorer.exe)
For years- I have been using a combination of Sonoff S31 Plugs and a Kasa HS300 Strip to monitor, and measure my power consumption.
This- allowed me to fully visualize my rack's power consumption, into a sankey diagram.
However- the form-factor of neither solution was ideal, and as my rack has become nearly full, I needed a better solution as PDUs filled with smart-plugs, isn't exactly ideal.
The issue is, there are not many 120v PDUs, for a 24U rack. As a matter of fact, there are not many zero-U PDUs at all for 24U racks.
But- I finally found an option, for a fully managed PDU, with INDIVIDUALLY metered and switched outlets, which CAN2 plug into a standard 120v outlet.
So... normally when you want faster networking, you can just upgrade to faster NICs (10g, 25, 40, 50, 100g.. etc...).
However, in the case of my Synology DS423+, you get, two 1G ethernet ports with no expansion.
You would think the best course of action, would be to leverage LACP- however that is NOT the case.
Enabling Multichannel can double the amount of bandwidth per client, where as LACP, depending on hashing, only gives benefit up to the performance of one port.
This post details how to enable multipath for SMB, and how to configure iSCSI on linux with multiple paths.