New performance patch available for testing on stable
The URL: http://apollo.backplane.com/FreeBSD4/ This patch reworks the pageout daemon and the buf_daemon. It is based on my previous patch but hopefully has the kinks worked out. The patch is for -stable only, I will have a -current patch tonight. The main change is to the buf_daemon. I ripped out the dynamic speedup / slowdown code because it just wasn't working. I replaced it with code to track the amount of I/O in-progress in order to be able to limit the flush rate based on that. I also ripped out the non-working hysteresis for waking up the buf_daemon and put in real hysteresis, and I ripped out the artificial limitations on the number of buffers that could be flushed per wakeup (just like I ripped out the maxlaunder limitation in the pageout daemon in the last patch), and instead flush until the low water mark is reached, using the runningbufspace (bytes in transit to the I/O device) to limit the flush rate. In my testing, these changes lead to much, much smoother operation in heavily loaded situations and also appears to improve the write rate: 4.2-STABLE: (dd'ing to a CCD stripe of two SCSI drives) serv02:/data1# dd if=/dev/zero of=test bs=32k count=49152 49152+0 records in 49152+0 records out 1610612736 bytes transferred in 34.506011 secs (46676295 bytes/sec) 4.2-STABLE + patches: (dd'ing to a CCD stripe of two SCSI drives) serv01:/data1# dd if=/dev/zero of=test bs=32k count=49152 49152+0 records in 49152+0 records out 1610612736 bytes transferred in 27.995698 secs (57530722 bytes/sec) The current patchset will pageout a little more then 4.2-RELEASE, but hopefully to the benefit of the system rather then the detriment. This is because I ripped out the two-pass inactive queue scan in the pageout daemon that was skipping dirty pages in the first pass (giving them too much priority) and replaced it with a one-pass scan. I am slowly making my way to per-(disk)-device I/O pipelining. At the moment the pipelining is system-wide. -Matt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: softupdates and /
: :So to keep Softupdates active on all boot ups, you have to re-execute :"tunefs -n enable /" or "tunefs -n enable /usr"? Otherwise its not active :on the partition? : :Jorge Once you turn softupdates on with tunefs, it's on for good. Of course, your kernel has to be compiled with the SOFTUPDATES option to actually run softupdates. -Matt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: softupdates and /
On 15-Dec-00 flag wrote: On Mon, 11 Dec 2000, Alfred Perlstein wrote: * [EMAIL PROTECTED] [EMAIL PROTECTED] [001211 11:50] wrote: Hi, is there anything wrong with enabling softupdates on my / partition? Yes and no. uhhhmanyway i'm unable to set softupdates on / cause tunefs needs a umounted partition and / is always mounted... Umm, if / is mounted read-only you can tunefs it. i.e., in single user mode after boot, or drop to single user and use 'mount -o ro -u /' then do the tunefs. However, it doesn't buy you much on /, and there is still some risk with it, so I'd not recommend it for /. how have you done it? thanks Paolo -- John Baldwin [EMAIL PROTECTED] -- http://www.FreeBSD.org/~jhb/ PGP Key: http://www.baldwin.cx/~john/pgpkey.asc "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: Dedicated disks (was: Dangerously Dedicated)
Is there anyone interested in rewriting that "fake" partition table, or is that requirement satisfied by the non-dedicated format? I actually like sysinstall, now that I am used to it, but it would be aesthetically more pleasing to be able to use the dedicated format. I am curious if there could be some improvement to that "fake" partition table so it would boot from any BIOS, even if it means we have to write some "proprietary" signatures there. Is it possible to modify it so it will trick the BIOS's that are currently failing to boot? On Fri, 15 Dec 2000, Greg Lehey wrote: On Friday, 15 December 2000 at 2:20:40 -0500, Mike Nowlin wrote: Does that mean that such BIOS's are proprietary in the sense that they don't recognize the dedicated format? There are times when the politically-correct of the world use the term "proprietary" when they actually mean "dumb" or "really badly designed". But yes, that's what it means... :) To be fair, the dedicated fake partition table format is a hack. It's too difficult to figure out what the real geometry is, so it invents one which should "do the job". Some BIOSes check the table and find it wanting. It's a grey area. Greg -- Finger [EMAIL PROTECTED] for PGP public key See complete headers for address and phone numbers To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Bridge + Firewall + Dummynet (all Working) + ... routing ????
Hello, I would like to know if anybody here did tried that before: I have a Bridge+Firewall+Dummynet box working wonderfully (4.2-Stable) between one of our networks and a CISCO router (a default gateway) to Internet. We have chosen the Bridge model because using the transparent bridge we didn't even change all "default gateway" in all boxes inside our Network. Is there any way (perhaps using Divert and Natd ?) to implement some basic static routes in this bridge ? The reason is I nowadays we have many IP classes and I feel it is a shame that one packet must go through the bridge to reach the Router, than come back through the bridge again to be delivered to an Access Server, for instance. The way I know to achieve that is changing the bridge's IP address and disabling the bridge code (assigning another IP address to the Router), but before I go this way, I'm trying to keep just one hop to get out of my Network ;-) Best Regards, Antonio Carlos Pina To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: Sorry! Unable to transfer
I got the same errors. On Fri, 15 Dec 2000, Nader Turki wrote: Hello all, I just got the FreeBSD 4.2-R ISO for the 2nd time from ftp://ftp.FreeBSD.org/pub/FreeBSD/releases/i386/ISO-IMAGES/4.2-install.iso While installing the XFree86 Distribution I get few errors that says: Unable to transfer the PC98-Servers/X9480 distribution from acd0c. Do you want to try to retrive again? Yes No I get the same message with thr following: PC98-Servers/X9EGC PC98-Servers/X9GAq PC98-Servers/X9GAN PC98-Servers/X9LPIw PC98-Servers/X9MGA PC98-Servers/X9NKV PC98-Servers/X9NS3 PC98-Servers/X9SPW PC98-Servers/X9SVG PC98-Servers/X9TGU PC98-Servers/X9WEP PC98-Servers/X9WS PC98-Servers/X9WSN and: Xlk98 X9set I just installed FreeBSD 4.1.1-R and didn't get those errors. I'm sure it's not from my CDROM. I mean is that somethin' normal? should i just ignore those files? Thank you, -Nader To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: Staying Stable
"Kartic Krishnamurthy" [EMAIL PROTECTED] writes: 1. What is the recommended frequency of cvsuping the source tree to stay stable with freebsd? If your machine is a production server: whenever we fix a bug that you know is or might affect your system, and the consequences of that bug are worse than the downtime necessary for upgrading. If it's just your home playbox: whenever you're bored and have nothing better to do. DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: Trap 12/Squid/Fbsd 4.1.1
Jesus Rodriguez [EMAIL PROTECTED] writes: Please, could some one give me a description of this fatal trap: ? Please refer to section 13.13 of the FAQ. DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: softupdates and /
John Baldwin [EMAIL PROTECTED] writes: Umm, if / is mounted read-only you can tunefs it. i.e., in single user mode after boot, or drop to single user and use 'mount -o ro -u /' then do the tunefs. However, it doesn't buy you much on /, and there is still some risk with it, so I'd not recommend it for /. On the contrary - if you have a fairly standard setup with /tmp on the root partition, you will benefit a lot from enabling softupdates on it. As to the risks, nowadays I'd say the odds of the softupdates code exhibiting a noticeable bug are about the same (or less) as those of a physical drive failure. YMMV. DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
No Subject
No Subject
Re: Dedicated disks (was: Dangerously Dedicated)
Glendon Gross: Is there anyone interested in rewriting that "fake" partition table, Please look at the thread with the same topic three weeks ago. I stated that it wouldn't be possible because there is a fundamental disagreement: BIOS standard demands that the first *sector* always remains reserved. However DD mode only reserves the first *block* and starts with the actual contents at the second block. This (intentionally) violates BIOS standards. Though many BIOSsen will happily accept it, some will get picky and refuse to boot a DD disk due to standards violation. If you would leave the first *sector* reserved with DD mode, you are basically left with the layout of non-DD mode, so this is pointless. or is that requirement satisfied by the non-dedicated format? Exactly. Non-DD follows the BIOS standards, so it will serve you fine, at the "expense" of losing some few kB of disk space. it would be aesthetically more pleasing to be able to use the dedicated format. Fully agree. However there are things in the BIOS architecture that don't care about aesthetics. After all, this is the PC world. Helge To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
