From owner-freebsd-hackers Sun May 28 20:39:11 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id UAA13675 for hackers-outgoing; Sun, 28 May 1995 20:39:11 -0700 Received: from haywire.DIALix.COM (peter@haywire.DIALix.COM [192.203.228.65]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id UAA13643 for ; Sun, 28 May 1995 20:37:51 -0700 Received: (from peter@localhost) by haywire.DIALix.COM (8.6.12/8.6.12/DIALix) id LAA03778; Mon, 29 May 1995 11:36:04 +0800 Date: Mon, 29 May 1995 11:36:03 +0800 (WST) From: Peter Wemm To: hackers@FreeBSD.org Subject: Re: aargh! How to handle 128 ttys on a system? In-Reply-To: <199505282352.QAA09627@ref.tfs.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: hackers-owner@FreeBSD.org Precedence: bulk On Sun, 28 May 1995, Julian Elischer wrote: > > and he's running on what is (to some extent) my machine I have when I'm > at home in PERTH.. :) > (I think it has a different motherboard) Yep. I actually moved the major components from my work machine into Julian's, because I went for a nostalgia trip and with SimCity under X. :-) Along the way, I got rather interested... :-) I've spent a lot of time in the past porting stuff from the *BSD systems into SVR4 rather than change OS's (because of the lack of drivers for various hardware in the free OS's). With the specialix driver basically working, that cures _the_ major obstacle for us. BTW: Julian: What came over you when you partitioned that drive? 31 Meg swap space? :-) ?? Sure makes it hard to get a crash dump of a machine with 32M in it. (I've since found about options "MAXMEM=31000") > DIALIX is one of the major ISP players in Perth and Australia.. > it'd be a nice reference :) > (You may remember that when i'm in AUS my address is julian@dialix.oz.au) NetBSD has done quite well for itself in this area.. There are a few few NetBSD based providers over here that spring to mind. > > just in case you are hanging back, not sure about this. here is the > > nic's data on peter ;) > Why would anyone be hanging back :) ? > I'm sure he exists (or that hologram drives pretty well) Heh. Tell that to the owner of the porsche that I bumped into last year... :-) Oh, also, I restored a '-p' option to tunefs that I had been missing from my SVR4 machines... On Julian's drive, I see: tunefs: maximum contiguous block count: (-a) 1 tunefs: rotational delay between contiguous blocks: (-d) 4 ms tunefs: maximum blocks per file in a cylinder group: (-e) 1024 tunefs: minimum percentage of free space: (-m) 10% tunefs: optimization preference: (-o) time I have a question: Are these appropriate for clustering? I remember with SunOS, rotdelay needed to be zero and maxcontig > 1 for clustering to kick in. Is this still the case with FreeBSD's clustering? Patch follows.. (It also fixes a documentation bug, -A was not in the manpage). Yes, I know the same info can be extracted with dumpfs, but I think this is far more convenient than trying to find which options are tuneable in the largeish dumpfs output. Cheers, -Peter *** tunefs.old.c Thu May 26 14:35:06 1994 --- tunefs.c Mon May 29 10:09:20 1995 *************** *** 117,122 **** --- 117,137 ---- Aflag++; continue; + case 'p': + warnx("maximum contiguous block count: (-a) %d", sblock.fs_maxcontig); + warnx("rotational delay between contiguous blocks: (-d) %d ms", sblock.fs_rotdelay); + warnx("maximum blocks per file in a cylinder group: (-e) %d", sblock.fs_maxbpg); + warnx("minimum percentage of free space: (-m) %d%%", sblock.fs_minfree); + warnx("optimization preference: (-o) %s", sblock.fs_optim == FS_OPTSPACE ? "space" : "time"); + if (sblock.fs_minfree >= MINFREE && + sblock.fs_optim == FS_OPTSPACE) + warnx(OPTWARN, "time", ">=", MINFREE); + if (sblock.fs_minfree < MINFREE && + sblock.fs_optim == FS_OPTTIME) + warnx(OPTWARN, "space", "<", MINFREE); + + continue; + case 'a': name = "maximum contiguous block count"; if (argc < 1) *************** *** 233,238 **** --- 248,254 ---- fprintf(stderr, "\t-e maximum blocks per file in a cylinder group\n"); fprintf(stderr, "\t-m minimum percentage of free space\n"); fprintf(stderr, "\t-o optimization preference (`space' or `time')\n"); + fprintf(stderr, "\t-p no change - just prints current tuneable settings\n"); exit(2); } *** tunefs.old.8 Wed Mar 15 17:10:47 1995 --- tunefs.8 Mon May 29 10:06:18 1995 *************** *** 39,48 **** --- 39,50 ---- .Nd tune up an existing file system .Sh SYNOPSIS .Nm tunefs + .Op Fl A .Op Fl a Ar maxcontig .Op Fl d Ar rotdelay .Op Fl e Ar maxbpg .Op Fl m Ar minfree + .Op Fl p .Bk -words .Op Fl o Ar optimize_preference .Ek *************** *** 54,59 **** --- 56,65 ---- The parameters which are to be changed are indicated by the flags given below: .Bl -tag -width Ds + .It Fl A + The file system has several backups of the super-block. Specifying + this option will cause all backups to be modified instead of only the + primary super-block. This is potentially dangerous - use with caution. .It Fl a Ar maxcontig This specifies the maximum number of contiguous blocks that will be laid out before forcing a rotational delay (see *************** *** 103,108 **** --- 109,120 ---- higher overhead for file writes. The kernel normally changes the preference automatically as the percent fragmentation changes on the file system. + .It Fl p + This option shows a summary of what the current tuneable settings + are on the selected file system. More detailed information can be + obtained in the + .Xr dumpfs 8 + manual page. .El .Sh SEE ALSO .Xr fs 5 ,