On 22 Feb 2006, at 9:47, Jorn Hass wrote: > > Hello decker, > > Wednesday, February 22, 2006, 2:14:00 AM, you wrote: > > > > Hellos, > > > I think this is the same patch yall are trying with, I couldn't > > verify since the bug/pr database was "busy" when I just tried to > > look. http://xmail.n3t.net/kern_event.c.patch > > Yep, that's exactly the same patch. > My box however, originally didn't last more than a day and then > bombed... Haven't stop/started since... > 5 days later: > Fri09AM 18:45.35 /var/MailRoot/bin/XMail > > lsof output: > > XMail 639 root cwd VDIR 4,27 512 2 / > XMail 639 root rtd VDIR 4,27 512 2 / > XMail 639 root txt VREG 4,37 323656 26940 > /var/MailRoot/bin/XMail XMail 639 root txt VREG > 4,27 142236 32165 /libexec/ld-elf.so.1 XMail 639 root > txt VREG 4,27 20228 51 /lib/libkvm.so.2 XMail > 639 root txt VREG 4,27 28644 50 > /lib/libcrypt.so.2 XMail 639 root txt VREG 4,31 > 102568 47181 /usr/lib/libc_r.so.5 XMail 639 root txt > VREG 4,31 834584 47354 /usr/lib/libstdc++.so.4 XMail > 639 root txt VREG 4,27 120004 52 /lib/libm.so.3 > XMail 639 root txt VREG 4,31 136020 47297 > /usr/lib/libpthread.so.1 XMail 639 root txt VREG > 4,27 884716 61 /lib/libc.so.5 XMail 639 root 0u > IPv4 0xc1e46de0 0t0 TCP mailbox:XMailCTL (LISTEN) XMail > 639 root 1u VCHR 2,2 0t0 22 /dev/null XMail > 639 root 2u VCHR 2,2 0t0 22 /dev/null > XMail 639 root 3u PIPE 0xc1d3d600 16384 > ->0xc1d3d6ac XMail 639 root 4u PIPE 0xc1d3d6ac > 0 ->0xc1d3d600 XMail 639 root 5u IPv4 0xc1e47000 > 0t0 TCP mailbox:pop3 (LISTEN) XMail 639 root 6u > IPv4 0xc1e47c24 0t0 TCP mailbox:smtp (LISTEN) XMail > 639 root 7u IPv4 0xc1e47a68 0t0 TCP localhost:finger > (LISTEN) XMail 639 root 8w VREG 4,27 0 > 32217 / (/dev/mirror/gm0s1a) > > Not a single KQUEUE in sight... :) > > David, just check if "lsof" gives you a similar output, or if there > are glaring differences. Maybe something is not quite right in the > garden of eden... > > And please note that I tied the finger port to localhost only... :P
Above lsof output is similar to what I get now. The patch certainly prevents kqueue entries but seems to destabilise the system. Since I'd been able to mostly get round the problem by restart of xmail at intervals of a couple of weeks or so I was then finding that once in a while there would be a kernel panic on stopping xmail. I'd decided it was a hardware related problem, either fault or incompatibility with FreeBSD > 4.7. Before moving xmail to the 586-133 I had it running for some months on a Celeron 2400 box with FreeBSD 4.10 -> 5.2.1 -> 5.3 (GENERIC kernels), I had no problem at all but on checking now and starting xmail I can see that within about 30 minutes lsof indicates the KQUEUE problem is present. More recently I'd seen mention of a disk/filesystem problem confirmed in FreeBSD 6, subsequently identified and supposedly fixed, and that bug had been present through 5.x but I've not followed this up as I'm in process of moving back onto NetBSD anyway. David - To unsubscribe from this list: send the line "unsubscribe xmail" in the body of a message to [EMAIL PROTECTED] For general help: send the line "help" in the body of a message to [EMAIL PROTECTED]
