Date: Sun, 19 Apr 2009 02:10:00 -0700 From: Doug Hardie <bc979@lafn.org> To: freebsd-stable Stable <freebsd-stable@freebsd.org> Subject: Re: FreeBSD 7.2-rc1 Message-ID: <40A886C4-9202-48F8-B111-F4C7F4E3EA11@lafn.org> In-Reply-To: <84888337-9024-4DA0-BF29-06FF98A49F15@lafn.org> References: <84888337-9024-4DA0-BF29-06FF98A49F15@lafn.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Apr 18, 2009, at 14:05, Doug Hardie wrote: > I have encountered a rather interesting issue trying to install > rc1. The system boots and then says there is no disk in the CD > drive. The rc1 disk1 downloaded fine and the checksums matched. > The CD will mount fine in other systems and can easily be read. I > then let 7.0 boot on the test system and mounted the 7.2 cd. It > mounts fine and I can read all the files ( well a few that I > tested). Looking through the 7.0 dmesg I find some rather > unexpected entries for the CD drive. ... Since I can't boot the CD, I did a source update. Unfortunately I seem to have downloaded one of the kernel modules while it was being updated since it would not compile. Since it was for hardware I don't have, I just commented it out and everything then built just fine. #device ath # Atheros pci/cardbus NIC's #device ath_hal # Atheros HAL (Hardware Access Layer) #device ath_rate_sample # SampleRate tx rate control for ath I usually run a stress test on new releases. Basically I open a ftp to a server (local LAN) and download gigs of data - more than the available free space on the drive. With 7.1, the download would hang about 50% of the time. There was free space remaining on the drive. The NIC was basically useless at that point. The only way to restore it was to reboot. The particular machine I am using right now only has an old decrepit rl NIC: rl0: <RealTek 8139 10/100BaseTX> port 0xe000-0xe0ff mem 0xdd104000-0xdd1040ff irq 12 at device 13.0 on pci0 rlphy0: <RealTek internal media interface> PHY 0 on miibus0 rlphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto rl0: Ethernet address: 00:d0:09:d8:8f:ff rl0: [ITHREAD] When running this test with 7.2-rc1 it always managed to make it to disk full and then terminate gracefully. Looks like the rl NICs are going to be usable. I had eliminated them from my production servers a long time ago. The only real issue I encountered was that the number of files you have to respond to in mergemaster continues to grow (perhaps exponentially). For my machine I maintain the source on thats no big issue. However, it does cause a lot of additional down time for the servers. I am going to have to dig through mergemaster to see if there is some way to tell it to automatically install the updates to specific directories (e.g., periodic, security, rc.d etc.). I never touch them and they contribute the majority of manual entries.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?40A886C4-9202-48F8-B111-F4C7F4E3EA11>