Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751782AbdIWVhv (ORCPT ); Sat, 23 Sep 2017 17:37:51 -0400 Received: from gherkin.frus.com ([192.158.254.49]:45272 "EHLO gherkin.frus.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750847AbdIWVhu (ORCPT ); Sat, 23 Sep 2017 17:37:50 -0400 X-Greylist: delayed 1728 seconds by postgrey-1.27 at vger.kernel.org; Sat, 23 Sep 2017 17:37:49 EDT Date: Sat, 23 Sep 2017 16:09:00 -0500 From: Bob Tracy To: linux-kernel@vger.kernel.org Subject: [INFO] "git ac --auto" for kernel repo Message-ID: <20170923210900.GA27116@gherkin.frus.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.23 (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1426 Lines: 29 A recent pull on Linus' kernel repo triggered a "git gc --auto" maintenance action that *finally* finished after beating my poor Alpha PWS 433au to death for 32 hours. Here's a snapshot of the ".git/objects/pack" directory for the repo: total 1366084 190535 12 drwxr-xr-x 2 root root 12288 Sep 23 15:31 . 190534 4 drwxr-xr-x 4 root root 4096 Sep 23 15:10 .. 191004 153720 -r--r--r-- 1 root root 157249268 Sep 23 15:10 pack-d11d36b9439b018666251bf101b920e27e3607e4.idx 190906 1212348 -r--r--r-- 1 root root 1240224629 Sep 23 15:08 pack-d11d36b9439b018666251bf101b920e27e3607e4.pack The "gc" action ran for more than a day before the temporary pack file showed up. When the smoke finally cleared about a half-hour ago, the other pack files had evidently been consolidated into (or replaced by) what you see above. Only reason I figured this might be of general interest to the community, I saw other people questioning how long the automatic "gc" action takes to run. Reasonable (for some definition of the word) machines were said to take a few hours in the worst case scenario of several loose objects in a repo the size of the one containing 2.6 and later kernel versions. Definitely not looking to make this maintenance action a regular thing. The system was barely usable during the time it was running, but only having 1.5 GB RAM (maxed-out) definitely has a lot to do with that :-). --Bob