From owner-freebsd-qa Tue Aug 21 14: 9:16 2001 Delivered-To: freebsd-qa@freebsd.org Received: from spoon.beta.com (spoon.beta.com [199.165.180.1]) by hub.freebsd.org (Postfix) with ESMTP id 4010537B408 for ; Tue, 21 Aug 2001 14:09:14 -0700 (PDT) (envelope-from mcgovern@spoon.beta.com) Received: from spoon.beta.com (mcgovern@localhost.beta.com [127.0.0.1]) by spoon.beta.com (8.11.3/8.11.3) with ESMTP id f7LL9Dr05047 for ; Tue, 21 Aug 2001 17:09:13 -0400 (EDT) (envelope-from mcgovern@spoon.beta.com) Message-Id: <200108212109.f7LL9Dr05047@spoon.beta.com> To: freebsd-qa@freebsd.org Subject: Sysinstall whine... Date: Tue, 21 Aug 2001 17:09:13 -0400 From: Brian McGovern Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On the packages menu, it appears the only way to get out of a category menu is by selecting 'Cancel'. Both Space and Return/Enter [de]select packages, and there is no "<<< Exit" menu item. This seems counter-intuitive. In addition, selecting "Cancel" does not undo the changes made on the menu, again being counter-intuitive. You may therefore wish to change the "Cancel" button to a "Back" or "Exit" button, which may give a user more of a clue about what its for. -Brian To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Tue Aug 21 14: 9:40 2001 Delivered-To: freebsd-qa@freebsd.org Received: from funnel.cisco.com (funnel.cisco.com [161.44.131.24]) by hub.freebsd.org (Postfix) with ESMTP id A556F37B407 for ; Tue, 21 Aug 2001 14:09:37 -0700 (PDT) (envelope-from bmcgover@bmcgover-pc.cisco.com) Received: from bmcgover-pc.cisco.com (root@bmcgover-pc.cisco.com [161.44.149.69]) by funnel.cisco.com (8.8.5-Cisco.1/8.6.5) with ESMTP id RAA10019 for ; Tue, 21 Aug 2001 17:09:36 -0400 (EDT) Received: from bmcgover-pc.cisco.com (bmcgover@localhost.cisco.com [127.0.0.1]) by bmcgover-pc.cisco.com (8.11.3/8.11.3) with ESMTP id f7LL6Ua02170 for ; Tue, 21 Aug 2001 17:06:31 -0400 (EDT) (envelope-from bmcgover@bmcgover-pc.cisco.com) Message-Id: <200108212106.f7LL6Ua02170@bmcgover-pc.cisco.com> To: freebsd-qa@freebsd.org Subject: Looking for second opinion on Vinum speed on 4.4 RC1... Date: Tue, 21 Aug 2001 17:06:30 -0400 From: Brian McGovern Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG I'm running a couple of mirrored disks under vinum. I've done the config, and started the second plex of the mirror to bring the drive up. After, oh, about 6 or so hours, the subdisk is 47% complete. systat -vmstat is showing the first drive in the mirror doing 2.53MB/s at 12% utilization, and the second drive doing 2.53MB/s at 87% utilization. The drives are ~80GB. The configuration is: drive a device /dev/ad2s1e drive b device /dev/ad3s1e volume vol0 plex name vol0.p0 org concat vol vol0 plex name vol0.p1 org concat vol vol0 sd name vol0.p0.s0 drive a plex vol0.p0 len 160071332s driveoffset 265s plexoff set 0s sd name vol0.p1.s0 drive b plex vol0.p1 len 160071332s driveoffset 265s plexoff set 0s Overall, I'm not complaining, as it all seems to be working the way I'd expect, and I have no crash dumps to share (a milestone in and of itself :) ). However, it just seems to be going really slow, as the 80GB SCSI disk set took like 30 minutes to come up. I really can't believe that the IDE disks are 20 times slower. So, if someone has a set of disks they can mirror, can they try it, and let me know the overall performance? MB/s or how long it takes to get two "N" GB drives in synch would be useful. The one thing that I can think of that may be biting me is that these drives are master/slave on the same controller... I've been too cheap to go buy another Promise controller to get a couple of extra busses... -Brian To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Thu Aug 23 9: 7:27 2001 Delivered-To: freebsd-qa@freebsd.org Received: from funnel.cisco.com (funnel.cisco.com [161.44.131.24]) by hub.freebsd.org (Postfix) with ESMTP id 2B6FA37B415; Thu, 23 Aug 2001 09:07:17 -0700 (PDT) (envelope-from bmcgover@bmcgover-pc.cisco.com) Received: from bmcgover-pc.cisco.com (root@bmcgover-pc.cisco.com [161.44.149.69]) by funnel.cisco.com (8.8.5-Cisco.1/8.6.5) with ESMTP id MAA18912; Thu, 23 Aug 2001 12:07:16 -0400 (EDT) Received: from bmcgover-pc.cisco.com (bmcgover@localhost.cisco.com [127.0.0.1]) by bmcgover-pc.cisco.com (8.11.3/8.11.3) with ESMTP id f7NG45j01008; Thu, 23 Aug 2001 12:04:05 -0400 (EDT) (envelope-from bmcgover@bmcgover-pc.cisco.com) Message-Id: <200108231604.f7NG45j01008@bmcgover-pc.cisco.com> To: freebsd-qa@freebsd.org Cc: freebsd-stable@freebsd.org Subject: top takes excessively long time to start on RC1.... Date: Thu, 23 Aug 2001 12:04:05 -0400 From: Brian McGovern Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG I've run in to a rather disturbing problem. When running a fresh install of 4.4RC1 (not an upgrade), I try to run top. It takes approximately 1 minute, 29 seconds to begin to display data. The machine is a pretty typical P3 with an Intel card, 3 IDE disks and 2 SCSI disks. systat -vmstat takes ~2 seconds to start (not bad). With multiple windows open, it appears that starting top isn't bogging the system down (it stays mostly idle). There are no more/less interrupts, top and systat don't show anything becoming a CPU hog, no drop in memory, etc. Subsequent invocations (2nd, 3rd, 4th, etc.) after stopping the first take as long. Parallel invocations take about 55 seconds to start displaying data. In looking at the top display, it appears (when running two) that one is in the RUN state, and the other in the select state. This may be a clue. -Brian To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Thu Aug 23 9:28:55 2001 Delivered-To: freebsd-qa@freebsd.org Received: from smtp1.sentex.ca (smtp1.sentex.ca [199.212.134.4]) by hub.freebsd.org (Postfix) with ESMTP id CFB6C37B421; Thu, 23 Aug 2001 09:28:47 -0700 (PDT) (envelope-from mike@sentex.net) Received: from simoeon.sentex.net (pyroxene.sentex.ca [199.212.134.18]) by smtp1.sentex.ca (8.11.6/8.11.6) with ESMTP id f7NGSiY43941; Thu, 23 Aug 2001 12:28:44 -0400 (EDT) (envelope-from mike@sentex.net) Message-Id: <5.1.0.14.0.20010823121939.03e928f0@marble.sentex.ca> X-Sender: mdtpop@marble.sentex.ca X-Mailer: QUALCOMM Windows Eudora Version 5.1 Date: Thu, 23 Aug 2001 12:23:15 -0400 To: Brian McGovern , freebsd-qa@FreeBSD.ORG From: Mike Tancsa Subject: Re: top takes excessively long time to start on RC1.... Cc: freebsd-stable@FreeBSD.ORG In-Reply-To: <200108231604.f7NG45j01008@bmcgover-pc.cisco.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG What if you start it with top -u ? Also, did you in make.conf adjust # top(1) uses a hash table for the user names. The size of this hash # can be tuned to match the number of local users. The table size should # be a prime number approximately twice as large as the number of lines in # /etc/passwd. The default number is 20011. # #TOP_TABLE_SIZE= 101 # At 12:04 PM 8/23/01 -0400, Brian McGovern wrote: >I've run in to a rather disturbing problem. > >When running a fresh install of 4.4RC1 (not an upgrade), I try to run top. It >takes approximately 1 minute, 29 seconds to begin to display data. > >The machine is a pretty typical P3 with an Intel card, 3 IDE disks and 2 >SCSI disks. > >systat -vmstat takes ~2 seconds to start (not bad). > >With multiple windows open, it appears that starting top isn't bogging >the system down (it stays mostly idle). There are no more/less interrupts, >top and systat don't show anything becoming a CPU hog, no drop in memory, etc. > >Subsequent invocations (2nd, 3rd, 4th, etc.) after stopping the first take as >long. Parallel invocations take about 55 seconds to start displaying data. > >In looking at the top display, it appears (when running two) that one is in >the RUN state, and the other in the select state. This may be a clue. > > -Brian > >To Unsubscribe: send mail to majordomo@FreeBSD.org >with "unsubscribe freebsd-stable" in the body of the message To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Thu Aug 23 9:31:35 2001 Delivered-To: freebsd-qa@freebsd.org Received: from ringworld.nanolink.com (discworld.nanolink.com [217.75.135.248]) by hub.freebsd.org (Postfix) with SMTP id 0BF1D37B406 for ; Thu, 23 Aug 2001 09:31:24 -0700 (PDT) (envelope-from roam@ringlet.net) Received: (qmail 1598 invoked by uid 1000); 23 Aug 2001 16:29:56 -0000 Date: Thu, 23 Aug 2001 19:29:56 +0300 From: Peter Pentchev To: Mike Tancsa Cc: Brian McGovern , freebsd-qa@FreeBSD.ORG, freebsd-stable@FreeBSD.ORG Subject: Re: top takes excessively long time to start on RC1.... Message-ID: <20010823192956.E503@ringworld.oblivion.bg> Mail-Followup-To: Mike Tancsa , Brian McGovern , freebsd-qa@FreeBSD.ORG, freebsd-stable@FreeBSD.ORG References: <200108231604.f7NG45j01008@bmcgover-pc.cisco.com> <5.1.0.14.0.20010823121939.03e928f0@marble.sentex.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <5.1.0.14.0.20010823121939.03e928f0@marble.sentex.ca>; from mike@sentex.net on Thu, Aug 23, 2001 at 12:23:15PM -0400 Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Thu, Aug 23, 2001 at 12:23:15PM -0400, Mike Tancsa wrote: > > What if you start it with top -u ? > > Also, did you in make.conf adjust > [snip] > #TOP_TABLE_SIZE= 101 He said it was a fresh 4.4RC1 install, no recompilation; so any make.conf adjustments would make no difference. G'luck, Peter -- I had to translate this sentence into English because I could not read the original Sanskrit. > At 12:04 PM 8/23/01 -0400, Brian McGovern wrote: > >I've run in to a rather disturbing problem. > > > >When running a fresh install of 4.4RC1 (not an upgrade), I try to run top. It > >takes approximately 1 minute, 29 seconds to begin to display data. > > > >The machine is a pretty typical P3 with an Intel card, 3 IDE disks and 2 > >SCSI disks. > > > >systat -vmstat takes ~2 seconds to start (not bad). > > > >With multiple windows open, it appears that starting top isn't bogging > >the system down (it stays mostly idle). There are no more/less interrupts, > >top and systat don't show anything becoming a CPU hog, no drop in memory, etc. > > > >Subsequent invocations (2nd, 3rd, 4th, etc.) after stopping the first take as > >long. Parallel invocations take about 55 seconds to start displaying data. > > > >In looking at the top display, it appears (when running two) that one is in > >the RUN state, and the other in the select state. This may be a clue. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Thu Aug 23 9:43:50 2001 Delivered-To: freebsd-qa@freebsd.org Received: from funnel.cisco.com (funnel.cisco.com [161.44.131.24]) by hub.freebsd.org (Postfix) with ESMTP id 806B737B40A; Thu, 23 Aug 2001 09:43:40 -0700 (PDT) (envelope-from bmcgover@bmcgover-pc.cisco.com) Received: from bmcgover-pc.cisco.com (root@bmcgover-pc.cisco.com [161.44.149.69]) by funnel.cisco.com (8.8.5-Cisco.1/8.6.5) with ESMTP id MAA23193; Thu, 23 Aug 2001 12:43:38 -0400 (EDT) Received: from bmcgover-pc.cisco.com (bmcgover@localhost.cisco.com [127.0.0.1]) by bmcgover-pc.cisco.com (8.11.3/8.11.3) with ESMTP id f7NGeRj01165; Thu, 23 Aug 2001 12:40:27 -0400 (EDT) (envelope-from bmcgover@bmcgover-pc.cisco.com) Message-Id: <200108231640.f7NGeRj01165@bmcgover-pc.cisco.com> To: Peter Pentchev Cc: Mike Tancsa , freebsd-qa@FreeBSD.ORG, freebsd-stable@FreeBSD.ORG, bmcgover@bmcgover-pc.cisco.com Subject: Re: top takes excessively long time to start on RC1.... In-reply-to: Your message of "Thu, 23 Aug 2001 19:29:56 +0300." <20010823192956.E503@ringworld.oblivion.bg> Date: Thu, 23 Aug 2001 12:40:27 -0400 From: Brian McGovern Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG top -u brings it down to about 40 seconds, on average (max was ~42, min ~38), 6 run sample. Peter is right. This is 'stock' from the 4.4RC1 ISO. I'll be able to collect some more data when I get home this evening from my other test machine. I'll be able to [dis]prove whether its machine specific, I'll be able to play with code and/or build it to profile it, and if I track it in to the kernel, I'll be able to slap a debugger on it. -Brian > On Thu, Aug 23, 2001 at 12:23:15PM -0400, Mike Tancsa wrote: > > > > What if you start it with top -u ? > > > > Also, did you in make.conf adjust > > > [snip] > > #TOP_TABLE_SIZE= 101 > > He said it was a fresh 4.4RC1 install, no recompilation; so any make.conf > adjustments would make no difference. > > G'luck, > Peter > > -- > I had to translate this sentence into English because I could not read the o riginal Sanskrit. > > > At 12:04 PM 8/23/01 -0400, Brian McGovern wrote: > > >I've run in to a rather disturbing problem. > > > > > >When running a fresh install of 4.4RC1 (not an upgrade), I try to run top . It > > >takes approximately 1 minute, 29 seconds to begin to display data. > > > > > >The machine is a pretty typical P3 with an Intel card, 3 IDE disks and 2 > > >SCSI disks. > > > > > >systat -vmstat takes ~2 seconds to start (not bad). > > > > > >With multiple windows open, it appears that starting top isn't bogging > > >the system down (it stays mostly idle). There are no more/less interrupts , > > >top and systat don't show anything becoming a CPU hog, no drop in memory, etc. > > > > > >Subsequent invocations (2nd, 3rd, 4th, etc.) after stopping the first tak e as > > >long. Parallel invocations take about 55 seconds to start displaying data . > > > > > >In looking at the top display, it appears (when running two) that one is in > > >the RUN state, and the other in the select state. This may be a clue. > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Thu Aug 23 10:10:26 2001 Delivered-To: freebsd-qa@freebsd.org Received: from smtp1.sentex.ca (smtp1.sentex.ca [199.212.134.4]) by hub.freebsd.org (Postfix) with ESMTP id 5904537B401 for ; Thu, 23 Aug 2001 10:10:16 -0700 (PDT) (envelope-from mike@sentex.net) Received: from simoeon.sentex.net (pyroxene.sentex.ca [199.212.134.18]) by smtp1.sentex.ca (8.11.6/8.11.6) with ESMTP id f7NHADf50380; Thu, 23 Aug 2001 13:10:13 -0400 (EDT) (envelope-from mike@sentex.net) Message-Id: <5.1.0.14.0.20010823125930.04c054c0@marble.sentex.ca> X-Sender: mdtpop@marble.sentex.ca X-Mailer: QUALCOMM Windows Eudora Version 5.1 Date: Thu, 23 Aug 2001 13:04:44 -0400 To: Brian McGovern From: Mike Tancsa Subject: sysinstall comments (was Re: top takes excessively long time to start on RC1.... Cc: freebsd-qa@FreeBSD.ORG In-Reply-To: <200108231640.f7NGeRj01165@bmcgover-pc.cisco.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG re: top, I am just in the process of doing the same and will test to see if I can duplicate the behaviour on my gear. On another note, one thing that seems to have changed with sysintall is the way the focus happens after certain submenu options that did not occur before. For example, on my machine with 2 IDE drives, ad0 and ad2. For some reason, ad2 gets listed as the "first" drive to choose from. This is somewhat counter intuitive. Then, after going through the fdisk submenu, just before going to disklabel focus remains on ad0. So if you just hit enter/OK, it deselects ad0, and you are put back in fdisk. This was not the behaviour before and is someone inconsistent with other submenu navigation. Note, this also happens after selecting the XFree options. ---Mike At 12:40 PM 8/23/01 -0400, Brian McGovern wrote: >top -u brings it down to about 40 seconds, on average (max was ~42, min >~38), 6 run sample. > >Peter is right. This is 'stock' from the 4.4RC1 ISO. > >I'll be able to collect some more data when I get home this evening from my >other test machine. I'll be able to [dis]prove whether its machine specific, >I'll be able to play with code and/or build it to profile it, and if I >track it >in to the kernel, I'll be able to slap a debugger on it. > > -Brian > > > On Thu, Aug 23, 2001 at 12:23:15PM -0400, Mike Tancsa wrote: > > > > > > What if you start it with top -u ? > > > > > > Also, did you in make.conf adjust > > > > > [snip] > > > #TOP_TABLE_SIZE= 101 > > > > He said it was a fresh 4.4RC1 install, no recompilation; so any make.conf > > adjustments would make no difference. > > > > G'luck, > > Peter > > > > -- > > I had to translate this sentence into English because I could not read > the o >riginal Sanskrit. > > > > > At 12:04 PM 8/23/01 -0400, Brian McGovern wrote: > > > >I've run in to a rather disturbing problem. > > > > > > > >When running a fresh install of 4.4RC1 (not an upgrade), I try to > run top >. It > > > >takes approximately 1 minute, 29 seconds to begin to display data. > > > > > > > >The machine is a pretty typical P3 with an Intel card, 3 IDE disks > and 2 > > > >SCSI disks. > > > > > > > >systat -vmstat takes ~2 seconds to start (not bad). > > > > > > > >With multiple windows open, it appears that starting top isn't bogging > > > >the system down (it stays mostly idle). There are no more/less > interrupts >, > > > >top and systat don't show anything becoming a CPU hog, no drop in > memory, > etc. > > > > > > > >Subsequent invocations (2nd, 3rd, 4th, etc.) after stopping the > first tak >e as > > > >long. Parallel invocations take about 55 seconds to start > displaying data >. > > > > > > > >In looking at the top display, it appears (when running two) that > one is >in > > > >the RUN state, and the other in the select state. This may be a clue. > > > >To Unsubscribe: send mail to majordomo@FreeBSD.org >with "unsubscribe freebsd-stable" in the body of the message To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Thu Aug 23 10:36:40 2001 Delivered-To: freebsd-qa@freebsd.org Received: from funnel.cisco.com (funnel.cisco.com [161.44.131.24]) by hub.freebsd.org (Postfix) with ESMTP id 23A2237B405 for ; Thu, 23 Aug 2001 10:36:29 -0700 (PDT) (envelope-from bmcgover@bmcgover-pc.cisco.com) Received: from bmcgover-pc.cisco.com (root@bmcgover-pc.cisco.com [161.44.149.69]) by funnel.cisco.com (8.8.5-Cisco.1/8.6.5) with ESMTP id NAA00614; Thu, 23 Aug 2001 13:36:23 -0400 (EDT) Received: from bmcgover-pc.cisco.com (bmcgover@localhost.cisco.com [127.0.0.1]) by bmcgover-pc.cisco.com (8.11.3/8.11.3) with ESMTP id f7NHXCj01343; Thu, 23 Aug 2001 13:33:12 -0400 (EDT) (envelope-from bmcgover@bmcgover-pc.cisco.com) Message-Id: <200108231733.f7NHXCj01343@bmcgover-pc.cisco.com> To: Mike Tancsa Cc: freebsd-qa@FreeBSD.ORG Subject: Re: sysinstall comments (was Re: top takes excessively long time to start on RC1.... In-reply-to: Your message of "Thu, 23 Aug 2001 13:04:44 EDT." <5.1.0.14.0.20010823125930.04c054c0@marble.sentex.ca> Date: Thu, 23 Aug 2001 13:33:12 -0400 From: Brian McGovern Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Mike, I'll check this out, too. I've only installed it on systems with either 1 drive, or a ton. You're right, though, ad0 did appear at the END of the list when I had 5 drives (3 IDE, 2 SCSI - which, BTW, is the one showing the top problem). I agree, the boot drive should appear consistently as the FIRST drive in the list. Wait until you get to the packages menu. :) -Brian > > re: top, I am just in the process of doing the same and will test to see if > I can duplicate the behaviour on my gear. > > On another note, one thing that seems to have changed with sysintall is the > way the focus happens after certain submenu options that did not occur > before. For example, on my machine with 2 IDE drives, ad0 and ad2. For > some reason, ad2 gets listed as the "first" drive to choose from. This is > somewhat counter intuitive. > > Then, after going through the fdisk submenu, just before going to disklabel > focus remains on ad0. So if you just hit enter/OK, it deselects ad0, and > you are put back in fdisk. This was not the behaviour before and is > someone inconsistent with other submenu navigation. Note, this also > happens after selecting the XFree options. > > ---Mike > > At 12:40 PM 8/23/01 -0400, Brian McGovern wrote: > >top -u brings it down to about 40 seconds, on average (max was ~42, min > >~38), 6 run sample. > > > >Peter is right. This is 'stock' from the 4.4RC1 ISO. > > > >I'll be able to collect some more data when I get home this evening from my > >other test machine. I'll be able to [dis]prove whether its machine specific , > >I'll be able to play with code and/or build it to profile it, and if I > >track it > >in to the kernel, I'll be able to slap a debugger on it. > > > > -Brian > > > > > On Thu, Aug 23, 2001 at 12:23:15PM -0400, Mike Tancsa wrote: > > > > > > > > What if you start it with top -u ? > > > > > > > > Also, did you in make.conf adjust > > > > > > > [snip] > > > > #TOP_TABLE_SIZE= 101 > > > > > > He said it was a fresh 4.4RC1 install, no recompilation; so any make.co nf > > > adjustments would make no difference. > > > > > > G'luck, > > > Peter > > > > > > -- > > > I had to translate this sentence into English because I could not read > > the o > >riginal Sanskrit. > > > > > > > At 12:04 PM 8/23/01 -0400, Brian McGovern wrote: > > > > >I've run in to a rather disturbing problem. > > > > > > > > > >When running a fresh install of 4.4RC1 (not an upgrade), I try to > > run top > >. It > > > > >takes approximately 1 minute, 29 seconds to begin to display data. > > > > > > > > > >The machine is a pretty typical P3 with an Intel card, 3 IDE disks > > and 2 > > > > >SCSI disks. > > > > > > > > > >systat -vmstat takes ~2 seconds to start (not bad). > > > > > > > > > >With multiple windows open, it appears that starting top isn't boggi ng > > > > >the system down (it stays mostly idle). There are no more/less > > interrupts > >, > > > > >top and systat don't show anything becoming a CPU hog, no drop in > > memory, > > etc. > > > > > > > > > >Subsequent invocations (2nd, 3rd, 4th, etc.) after stopping the > > first tak > >e as > > > > >long. Parallel invocations take about 55 seconds to start > > displaying data > >. > > > > > > > > > >In looking at the top display, it appears (when running two) that > > one is > >in > > > > >the RUN state, and the other in the select state. This may be a clue . > > > > > > >To Unsubscribe: send mail to majordomo@FreeBSD.org > >with "unsubscribe freebsd-stable" in the body of the message > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Thu Aug 23 11:45:39 2001 Delivered-To: freebsd-qa@freebsd.org Received: from guru.mired.org (okc-94-248-46.mmcable.com [24.94.248.46]) by hub.freebsd.org (Postfix) with SMTP id 8FCF337B403 for ; Thu, 23 Aug 2001 11:45:33 -0700 (PDT) (envelope-from mwm@mired.org) Received: (qmail 81152 invoked by uid 100); 23 Aug 2001 18:45:23 -0000 From: Mike Meyer MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <15237.20291.380638.888721@guru.mired.org> Date: Thu, 23 Aug 2001 13:45:23 -0500 To: Brian McGovern Cc: freebsd-qa@freebsd.org, freebsd-stable@freebsd.org Subject: Re: top takes excessively long time to start on RC1.... In-Reply-To: <200108231604.f7NG45j01008@bmcgover-pc.cisco.com> References: <200108231604.f7NG45j01008@bmcgover-pc.cisco.com> X-Mailer: VM 6.90 under 21.1 (patch 14) "Cuyahoga Valley" XEmacs Lucid X-face: "5Mnwy%?j>IIV\)A=):rjWL~NB2aH[}Yq8Z=u~vJ`"(,&SiLvbbz2W`;h9L,Yg`+vb1>RG% *h+%X^n0EZd>TM8_IB;a8F?(Fb"lw'IgCoyM.[Lg#r\ Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Brian McGovern types: > I've run in to a rather disturbing problem. > > When running a fresh install of 4.4RC1 (not an upgrade), I try to run top. It > takes approximately 1 minute, 29 seconds to begin to display data. FWIW, I've got a fresh install from the 4.4RC1 ISO. top displays the initial process list instantly, and starts displaying and updating system information in ~2 seconds. > systat -vmstat takes ~2 seconds to start (not bad). More like ~4 here. > The machine is a pretty typical P3 with an Intel card, 3 IDE disks and 2 > SCSI disks. Mine's a K6 on a VIA chipset, with 2 IDE disks and 2 IDE CDROMS, but only one of the disks is actually used. > With multiple windows open, it appears that starting top isn't bogging > the system down (it stays mostly idle). There are no more/less interrupts, > top and systat don't show anything becoming a CPU hog, no drop in memory, etc. I don't have X installed on the machine. You might try testing things from a console. > In looking at the top display, it appears (when running two) that one is in > the RUN state, and the other in the select state. This may be a clue. Mine does the same. When top collects the process information to display, it's running, hence it's going to be in the RUN state. When it's not updating, it's sitting waiting for a command or the timer to run out - which is what select is used for. So that seems natural as well. http://www.mired.org/home/mwm/ Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Thu Aug 23 11:57:43 2001 Delivered-To: freebsd-qa@freebsd.org Received: from funnel.cisco.com (funnel.cisco.com [161.44.131.24]) by hub.freebsd.org (Postfix) with ESMTP id 068F537B40A for ; Thu, 23 Aug 2001 11:57:41 -0700 (PDT) (envelope-from bmcgover@bmcgover-pc.cisco.com) Received: from bmcgover-pc.cisco.com (root@bmcgover-pc.cisco.com [161.44.149.69]) by funnel.cisco.com (8.8.5-Cisco.1/8.6.5) with ESMTP id OAA13173 for ; Thu, 23 Aug 2001 14:57:38 -0400 (EDT) Received: from bmcgover-pc.cisco.com (bmcgover@localhost.cisco.com [127.0.0.1]) by bmcgover-pc.cisco.com (8.11.3/8.11.3) with ESMTP id f7NIsRj34306 for ; Thu, 23 Aug 2001 14:54:27 -0400 (EDT) (envelope-from bmcgover@bmcgover-pc.cisco.com) Message-Id: <200108231854.f7NIsRj34306@bmcgover-pc.cisco.com> To: freebsd-qa@freebsd.org Subject: RC 2 due out soon? Schedule? Date: Thu, 23 Aug 2001 14:54:27 -0400 From: Brian McGovern Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Are there any plans to roll an RC2 any time soon? For that matter, has whomever is doing the Release Engineering in Jordan's wake put out a schedule of RCs and a release target for 4.4? I'm just curious so I can schedule my time to fall in to the pattern of RCs, whether I should be downloading daily -STABLEs to try to keep up to date, and whether 4.4 is going to slip out when I'm not looking... :) -Brian To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Thu Aug 23 13:14:10 2001 Delivered-To: freebsd-qa@freebsd.org Received: from obsecurity.dyndns.org (adsl-63-207-60-54.dsl.lsan03.pacbell.net [63.207.60.54]) by hub.freebsd.org (Postfix) with ESMTP id 771AD37B401; Thu, 23 Aug 2001 13:14:04 -0700 (PDT) (envelope-from kris@obsecurity.org) Received: by obsecurity.dyndns.org (Postfix, from userid 1000) id C4B6F66CD2; Thu, 23 Aug 2001 13:14:03 -0700 (PDT) Date: Thu, 23 Aug 2001 13:14:03 -0700 From: Kris Kennaway To: Brian McGovern Cc: Peter Pentchev , Mike Tancsa , freebsd-qa@FreeBSD.ORG, freebsd-stable@FreeBSD.ORG, bmcgover@bmcgover-pc.cisco.com Subject: Re: top takes excessively long time to start on RC1.... Message-ID: <20010823131403.B99242@xor.obsecurity.org> References: <20010823192956.E503@ringworld.oblivion.bg> <200108231640.f7NGeRj01165@bmcgover-pc.cisco.com> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-md5; protocol="application/pgp-signature"; boundary="aM3YZ0Iwxop3KEKx" Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <200108231640.f7NGeRj01165@bmcgover-pc.cisco.com>; from bmcgover@cisco.com on Thu, Aug 23, 2001 at 12:40:27PM -0400 Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG --aM3YZ0Iwxop3KEKx Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Aug 23, 2001 at 12:40:27PM -0400, Brian McGovern wrote: > top -u brings it down to about 40 seconds, on average (max was ~42, min > ~38), 6 run sample. >=20 > Peter is right. This is 'stock' from the 4.4RC1 ISO.=20 >=20 > I'll be able to collect some more data when I get home this evening from = my > other test machine. I'll be able to [dis]prove whether its machine specif= ic, > I'll be able to play with code and/or build it to profile it, and if I tr= ack it > in to the kernel, I'll be able to slap a debugger on it. If you can figure out what's going on or how to repeat it on other machines, please do. I made a change to the signal handling in top a week or so ago which should have been safe, and no-one else has reported any problems, but there's always a chance. Kris --aM3YZ0Iwxop3KEKx Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (FreeBSD) Comment: For info see http://www.gnupg.org iD8DBQE7hWQLWry0BWjoQKURAqIOAKCC4zoQuM0HvyW8Jb2CbUmkYBtrTACdFTBS uHseGr1c7/ZaLibuB5EPpHI= =Dnvo -----END PGP SIGNATURE----- --aM3YZ0Iwxop3KEKx-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message From owner-freebsd-qa Thu Aug 23 17:33:30 2001 Delivered-To: freebsd-qa@freebsd.org Received: from spoon.beta.com (spoon.beta.com [199.165.180.1]) by hub.freebsd.org (Postfix) with ESMTP id 1657F37B40B for ; Thu, 23 Aug 2001 17:33:28 -0700 (PDT) (envelope-from mcgovern@spoon.beta.com) Received: from spoon.beta.com (mcgovern@localhost.beta.com [127.0.0.1]) by spoon.beta.com (8.11.3/8.11.3) with ESMTP id f7O0XQr11414 for ; Thu, 23 Aug 2001 20:33:27 -0400 (EDT) (envelope-from mcgovern@spoon.beta.com) Message-Id: <200108240033.f7O0XQr11414@spoon.beta.com> To: freebsd-qa@freebsd.org Subject: More on top taking forever to start Date: Thu, 23 Aug 2001 20:33:26 -0400 From: Brian McGovern Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Ok. It does not occur on the home 4.4RC1 system. So, it should probably be something to do with the specific machine. I did notice one thing that may be of use. On my work system, irq 5, as per systat -vmstat, is labeled "mux". On the home system, irq 5 isn't even used. I've proven its not network related. Both servers have Intel cards in them. The home system worked fine over the net and on the console. Top didn't work either on the console or over the network at the office. When I ran top at the work system, systat -vmstat didn't show much system activity until it started to display, about when I saw 508 interrupts on the mux irq. When the system is idle, I see 2-3 mux interrupts every 5-10 seconds. So what happens is this.... Low "mux" interrupts. top is started, and a few more show up (although its within the background noise). Then, when top starts displaying, there is a sudden surge. I'll take a look and see if there are any serious interrupt conflicts, although I expect there aren't any. I'll also try to see if I can link them back to top. -Brian - --f7NMot414166.998607055/sj-msg-av-2.cisco.com-- ------- End of Forwarded Message To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message