FreeBSD MySQL still WAY slower than Linux
On Mon Jul 20, (I was reading the web archive) Mark Kirkwood wrote: With respect to Mysql performance, I would suspect threading or threading/kernel interaction as the culprit. (That reminds me, I don't recall seeing the original poster re-doing the tests with 6.0-CURRENT - that would be interesting). Sorry for not getting back sooner. I posted these results and anything else that went with this thread to the freebsd-performance mailing list. -current proved a slightly better performer for us, but not enough to bring it even close to the performance we get with gentoo. Off the top of my head the select key benchmarks were ROUGHLY: FreeBSD 5.various + MySQL 4.various: ROUGHLY 16k queries/second FreeBSD 6.0-current + MySQL 4.1.12 : ROUGHLY 18k queries/second Gentoo-something + MySQL 4.1.something : ROUGHLY 30k queries/second. The exact figures were posted to performance- (6.0 tests) and stable- (gentoo vs. freebsd 5.x). Good luck to anyone who can manage to get MySQL to perform as well on FreeBSD as Linux, we've been right out of luck so far and that's really not advocating FreeBSD very well, which is what I'd rather be doing. Steve ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: FreeBSD MySQL still WAY slower than Linux
I've posted a longer reply, with a trimmed cc list on the -performance mailing list if anyone is still interested and I'll leave it off -stable as it's probably become somewhat off topic now. Sadly, I don't think simply having an async FS is going to solve our problem though. :( Ta, Steve Roome For refernce: On Fri, Jun 17, 2005 at 09:28:54AM -0400, David Sze wrote: At 05:15 PM 16/06/2005 +0100, Steve Roome wrote this to All: It turns out that the problem was the same thing everyone usually points the finger at, but no one actually mentioned this time: Linux mounts its partitions async by default. I don't have the exact numbers in front of me right now, but these were the ballpark figures (I'm not going to separate out results for all of the different threading libraries for FreeBSD because the deltas weren't huge): ... ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: FreeBSD MySQL still WAY slower than Linux
Thank you all for your suggestions on this thread, here's a brief breakdown of most of the ideas from people: Billy Newsom: COMPILER, DISK, MYSQLVERSION Daniel Eischen: +/-HTT, Thread scopes Greg Lehey: MALLOC Guy Helmer: PREEMPTIVE, vfs.read_max Jon Dama: David Xu's Thrds, Ptmalloc, cpu affinity, sched+hwcacheing Kris Kennaway: +/-HTT Robert Watson: Thread scopes, LIBTHR/Linuxthr on 5 and 6?, LOCKING, HTT, SMP/UP Thomas Hurst: FreeBSD-current, Don't overload mysql! Vladimir Chukharev: COMPILEOPTS, TABLETYPES Xin Li: PROFILE, HTT insignificant The bad news is that I've not managed to get very far at all lately as MySQL has been crashing too much to even stop and test stuff elsewhere. The good news though, is that the Mysql folks have agreed to setup tests to profile mysql on identical hardware running FreeBSD and Linux with an aim to find out exactly where the problem really is. They reckon they'll spend at least two weeks trying to find out why Linux is so much faster - if they do this right I'd be surprised if we can't improve things quite a lot. Also, it'd be good if any of you who still have an interest in this could add any ideas or suggestions that may help *them* with their testing. If so, just get in touch with me asap before they get too stuck into anything that might prove fruitless. Here's hoping we can get MySQL running as well on FreeBSD as it ought to. Thanks again everyone, Steve Roome ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
FreeBSD MySQL still WAY slower than Linux
select_index10 21 3 15724.96
# FreeBSD 5.4-STABLE SMP, Hand-built 4.1.12, libkse system scope (dynamic)
[EMAIL PROTECTED] 7 0 % foreach f (1 2 3 4 5)
{/data/supersmack-1.3/bin/super-smack select-key.smack 50 1000|grep
select_index}
select_index10 22 0 15792.41
select_index10 8 0 15718.67
select_index10 4 0 15837.49
select_index10 5 0 15834.15
select_index10 3 0 15892.31
# FreeBSD 5.4-STABLE SMP, Hand-built 4.1.12, LinuxThreads
[EMAIL PROTECTED] 3 0 % foreach f (1 2 3 4 5)
{/data/supersmack-1.3/bin/super-smack select-key.smack 50 1000|grep
select_index}
select_index10 2 0 17709.35
select_index10 2 0 17701.14
select_index10 1 0 17758.04
select_index10 1 0 17829.17
select_index10 9 0 17557.00
# Gentoo Linux 2005.0, Hand-built 4.1.12, NPTL
[EMAIL PROTECTED] 1 0 % foreach f (1 2 3 4 5)
{/data/supersmack-1.3/bin/super-smack select-key.smack 50 1000|grep
select_index}
select_index10 0 0 34420.06
select_index10 2 0 34034.52
select_index10 4 0 33236.42
select_index10 3 0 33210.14
select_index10 1 0 34610.75
Thanks in advance for anyone that has a clue on this, and has anyone
figured out why FreeBSD is just so amazingly slow compared to Linux.
(That's not meant as flamebait, it just is.)
Steve Roome
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Makeworld is dying...
On Sun, Sep 17, 2000 at 04:00:06PM +0930, Greg Lehey wrote: What you *should* do before sending out a reply like this is to check whether it's really in the FAQ or not. It is (http://www.freebsd.org/FAQ/troubleshoot.html#AEN1570): There's a lot of people who are quick to step up to the podium and proclaim that all these signal elevens are caused by bad RAM, or bad RAM settings. It *could* be a pointer arithmetic problem. The FAQ ought to comment on this, and from recent personal experience, the number of people who respond to these just quoting the FAQ gets more annoying each time, mainly because it's something that doesn't need to be said on the mailing lists... again. In this case, it is possibly bad hardware, so maybes it's a fair call, then again it can be caused by, e.g. freaky optimizations in gcc... So could we change the text (something like, but better worded than the following) in the FAQ, e.g. : Q: My programs occasionally die with Signal 11 ( or 10 ). A: Signal 11 errors are caused when your process has attempted to access memory which the operating system has not granted it access to. This could be caused by a number of different circumstances : a) Most likely, if you're developing it yourself it's buggy code. (We've all been there!) b) If it's a problem with part of the base FreeBSD system, it might be buggy code, but more often than not these problems are found long before us general FAQ readers get to use these bits of code. If these problems are only affecting you, it's probably bad hardware. In the case of a) you can use a debugger and find the point in the program which is attempting to access a bogus address and then fix it. [ you probably already know this if you're a programmer! ] In the case of b) You need to verify the settings on your motherboard. Checking for hardware you might be running slightly out of spec, too fast, or mismatched hardware. Often setting memory wait states too short will trigger random signal 11's. An overclocked CPU will possibly also exhibit strange or similar symptoms. Try running some memory testing programs, or do a make buildworld if you have the full source available for FreeBSD (after a few successful buildworlds it's probably safe to say the hardware is okay.). See the SIG11 FAQ (LINK) for more information. That's my idea for a rough draft anyway. I'm clearly illiterate though, please don't flame me for that! Steve, To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: The kernel freeze when initializing my ELSA MicroLink ISDN/PCI ...
On Fri, Aug 18, 2000 at 08:23:17PM +0200, Matthieu Pasini wrote: Hello , I compiled the kernel with the following code inside : options ELSA_QS1PCI device isic0 pseudo-device "i4bq921" pseudo-device "i4bq931" pseudo-device "i4b" pseudo-device "i4btrc" 4 pseudo-device "i4bctl" pseudo-device "i4brbch" 4 pseudo-device "i4btel" 4 pseudo-device "i4bipr" 4 pseudo-device "i4bisppp" 4 pseudo-device sppp4 options IPR_VJ ** And the kernel freeze when it reach the initialization of the card ..., it says me : *** isic0: ELSA Microlink ISDN/PCI port . isic0: Error, IPAC version 2 unknow Fatal trap 12 : page fault while in kernel mode ... *** And it asks me for reboot. The problem is the same in 4.0-stable and in 4.1-stable. If anyone have had the same problem as me or if anyone know more about what's happening, please mail me ... Look for patches on the mailing list. There's some minor changes that need to be done in /sys/i4b/layer1 (they're very minor really!), I can't be sure if they'll get rid of the panic, but they'll sort out the IPAC version 2 unknown, and I beleive they stop the panic too. I spend some of these weekend helping someone out with this problem, and frankly I've seen the patches, applied the patches and still can't quite figure out why no-one has merged them in. So if you get the patches, send-pr the changes and hopefully someone will commit them to the tree, or alternatively, if it's a new -stable you might want to get an easier set of patches from someone who's managed to get it into the tree and has access to the box still. Steve P.S. I've cc'd the guy who has a copy of the patch, so if you can't find it ask him. Sorry to hassle you Simon. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: make buildworld: building strip fails.
On Sun, May 28, 2000 at 09:55:21AM -0700, Cy Schubert - ITSD Open Systems Group wrote: Same here. I did the same using a CVSup from Saturday May 27 @ 04:15 PDT on four machines. Everything went smooth. My only complaint is that FreeBSD is getting too big. On my P120 at home 3.4 used to buildworld in 4.5 hours now with 4.0 it took 7.75 hours, while my desktop system and Kerberos servers (333 MHz PII's) at work a 3.4 buildworld took 1.5 hours while building 4.0 world now took 2.25 hours. But I guess that's the price of progress. This is random guess that I can't check up on at the moment because I don't have a 4.0 system at work, but I kernel compiles don't use -pipe by default on 4.0. This of course may be due to them not being needed, so take this with a pinch of salt. Of course, on the other hand it might speed up your compiles a lot if my hunch is right. But it is just a hunch. Steve To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: 4-stable won't boot
On Wed, May 24, 2000 at 08:35:55PM -0700, Mike Smith wrote: Hello, I just did a cvsup to RELENG_4 about noon EST today (Wed), built world, installed a new kernel, and rebooted. now, when I boot, I get the following: ... Mounting root from ufs:/dev/ad0s3a no such device 'ad' Sounds like you left the 'ad' device out of your new kernel config. Bad idea. 8) For what it's worth, I had similar problems recently as well, although I was getting ata_master: timeout waiting for command (from memory), however I did have the ad device, and the problem seemed to be related to some sort of conflict that appeared only when I had one of the following also in the kernel : options AUTO_EOI_1 options AUTO_EOI_2 device pcm device pcf I've not narrowed it down much yet, but suffice it to say that I got almost the same problem described above, but which automagically fixed itself once I took out the above 4 lines. I doubt it's anything to do with this, but I thought I'd mention it in case someone following this thread knows why this might happen. I will look into it further once I've got the time to recompile another 20+ kernels and test them. Steve To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: Transparent proxies and fetch
On Thu, May 25, 2000 at 11:00:18AM +1000, Joe Shevland wrote: when building ports. Someone wrote in asking whether our ISP has implemented a transparent proxy. Unfortunately I culled my mail folder so I can't respond directly, but the poster was on the money (the ISP grabs any port 80 requests and stuffs them into the proxy instead). My question now is how do I work around this issue? I've tried setting the HTTP_PROXY variable but this just makes the 'make install' of the ports fail very quickly: In my limited experience calling the ISP and talking to them about it can work. I don't know if it will help, but you might be able to get them to remove the divert for you. The ISP I was with did that for me, on the other hand, you could always change ISP to someone who lets the customer decide when to use a proxy. Then again, someone might have a better solution, but IMHO I think it's quite rude of ISP's to divert your traffic without letting you know about it, imagine how you'd feel if they started diverting all your outgoing port23 connections and archiving everything that went down that line. Others may feel differently about it though! And it's becoming far too standard a practice now, so maybe we're supposed to move with the times and accept it? I dunno! Steve Roome To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
