[Bug 286591] PKG index for ARM64 seams do be out of date
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Fri, 16 May 2025 01:11:57 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=286591 Mark Millard <marklmi26-fbsd@yahoo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |marklmi26-fbsd@yahoo.com --- Comment #4 from Mark Millard <marklmi26-fbsd@yahoo.com> --- Looking at https://pkg-status.freebsd.org/builds?type=package&all=1 with appropriate searches and sorting shows: 142arm64's quarterly started building on Sat, 10 May 2025 03:10:18 GMT and still has 6716 to go, having built 28212 so far in 141:14:58 . The prior analogous, but much smaller (incemental) build, was for 141arm64 and started on Sat, 29 Mar 2025 05:35:24 GMT taking 50:28:17 . One issue is that as of pkg 2.1.0+ being involved, the (nearly) from-scratch builds take somewhat longer compared to when pkg 2.0.6 or before was involved. The arm64/armv7 systems (ampere*) are not the faster systems. Another is that ampere1 cycles through building 4 distinct build types (arm64/armv7 combined with 14.2/13.4), unlike amd64 systems that build just one type each and the next build is of the same type. (There are far more amd64/i386 systems.) A result of both of the above is that a build cycle of the 4 builds can take much of a month to complete and start the next cycle, at least when the cycle has large numbers of build to do for each type. No so long when they happen to be small incremental builds instead. (Quarterly gets such smaller builds far more often.) Another issue is that ampere1 has been crashing a lot in recent times. (Cause unknown as far as I know.) There were 4 crashes in April and 1 sigterm so far in May. The crashes insert delays. Do not expect arm64 and amd64 to be synchronized. If the 6716 finish without a crash, there should finally be a 142arm64 quarterly update, one that would include pkg 2.1.2 . -- You are receiving this mail because: You are on the CC list for the bug.