Re: trouble shooting samba performance
worms wrote: Hello, Can anyone point me to a guide on trouble shooting samba performance or a quick list of common issues I can check for. I've been googling but haven't found much that referenced a recent version of FreeBSD and Samba. I am running FreeBSD 7.0 on VmWare Server 2.0 hosted on a CentOS Linux box. I've used nttcp and iozone to verify that the performance of the FreeBSD 7.0 VM is good and am able to copy 6.5GB on disk from one location to another in about 2.5 minutes. Copying to a Windows 2003 server yields a 45 minute transfer time. Copying to a Windows XP workstation yields a 4 minute transfer time Summary: FreeBSD 7.0 --> FreeBSD 7.0 -- 2.5 minutes ( copying on disk ) FreeBSD 7.0 --> WinXP -- 4 minutes ( samba ) FreeBSD 7.0 -> Windows 2003 -- 45 minutes ( samba ) WinXP --> Windows 2003 -- 5 minutes Both windows machines are on the same domain. I've used samba for a number of years and this is the first time I've ran into a problem such as this. So if someone could give me a good starting point on how to troubleshoot this I'd appreciate it. Thanks --Lance The numbers indicate something i've seen several times but there has been different answers to it. Sometimes it was the indication that the RAID on the Windows2003 server was misconfigured or running in DEGRADED mode. However i've also seen issues regarding TCP/IP settings. Often it can be a good exercise to try to toggle this sysctl before transfer (net.inet.tcp.delayed_ack). Of course these are just clues and not real answers. -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Your suggestions about this Dell configuration?
VeeJay wrote: Hello Frank Really good points. I am really glad to have your thoughts. Regarding your questions and comments, I have given some answers and a couple of questions in *RED* colour. Please comment if you happen to manage some time during weekend, Thanks! *Please continue...* On Thu, May 22, 2008 at 4:01 AM, Frank Shute <[EMAIL PROTECTED]> wrote: On Wed, May 21, 2008 at 08:49:51AM +0200, VeeJay wrote: Hello friends, My employer is buying this Dell server and I would like to have your opinion about the configuration. Requirements are: 2 Websites with 3-4 million hits per month with video ads. If it's "3-4 million hits per month" as you've stated twice now, then your hardware is complete overkill. So I'll assume you mean 3-4 million hits a day for each site. *No, its 3-4 million each site per month and we are having problem. Because, either Apache or MySQL stops responding. I have following settings as Performance:* ** *# = # Performance settings # = Timeout 300 KeepAlive On MaxKeepAliveRequests 100 KeepAliveTimeout 15 MinSpareServers 5 MaxSpareServers 10 StartServers 5 MaxClients 256 MaxRequestsPerChild 0* Operating System: *FreeBSD AMD647-STABBLE* I'd use 7.0-RELEASE. Database: *PHP+MySQL with Apache* No problem. You should use Apache 2.*. *We will use Apache 2.** Server Configuration: *PowerEdge? 6850 SCSI* Dual Core Intel(R) Xeon(R) Processor 7130M, 3.2GHz, 8MB L3 Cache, 800Mhz FSB 1x Additional Dual Core Intel(R) Xeon(R) Processor 7130M, 3.2GHz, 8MB L3 Cache, 800MHz FSB Slow FSB. I suppose they hope you hit the cache. Shouldn't matter because your server is more likely to be disk bound rather than bus bound. Changed Processor to: *PE 2950 III Quad Core Xeon X5450 (3.0GHz, 2x6MB, 1333MHz FSB)* *what do you think about E5450?* 16GB 400MHz Dual Rank DDR2 Memory (8X2GB) Slow memory, to match the slow FSB :) But you've got >250MB per hit. So use the excess to cache frequently accessed content. *We have changed it to:* *16GB (8x2GB Dual Rank DIMMs) 667MHz FBD* C5 Drives attached to embedded PERC4ei, RAID 10 PERC 4/DC RAID controller (128MB cache) (1 intern and 1 extern Channel) (Should I use controller with Both Internal or Both External Channel? What they do?) Supported according to a quick Google search. 5 x 146GB SCSI Ultra320 (15000rpm) 1'' 80 pin harddrives "No name" or a brand? *We have changed the disks to :* *6 x 450GB SAS 15k 3.5" HD Hot Plug, (Hitachi Japan)* We have the PERC 4e/Di and that works wonderfully (256 MB battery backed cache). One hint is that the PERC 4e/Di (and possibly the entire series) does not do correct RAID 1 + 0 (mirrored then striped) but instead does RAID 1 and concatenates those mirrors. Chassis with support for 3.5'' SCSI Hard Drives Dell Remote Access Card 4 SERVER MANAGEMENT CARD Don't know if this will work. Most guys use a serial console/ssh for management. (I will have hot swappable drives & chassis) Get it if your server is going to be remote. It lets you mount CD-ROM disks and ISOs, Floppy images and gives you real keyboard/mouse/video display of the server. It also lets you power up/down/reboot the server remotely. A necessity to do firmware/BIOS upgrades. Serial console/ssh only lets you work with an already working system. The DRAC lets you do remote installs/reinstall/upgrades. Thank you in advance. The performance of this hardware will depend on what *sort* of hits you get. Are a lot of them just for the homepage? Then just cache it. Is it static content? *No, its dynamic contents, data is coming form Database.* If you're getting lots of ad-hoc database queries and fetches/writes from/to disk, then your disks could get a thrashing. How big's your database? Being read from more than written to? How precious is the data? *More than 20 million records and more than 1000 Tables.* *And of course, data is always preciouse. :)* How many of these hits are reading video ads? All of them? How many KBs are these awful ads? *50% of users are going to see the Video Ads.* ** *Size would vary between 100KB to 2MB. * What bandwidth do you have to these servers? *100 Mbps* How you are going to get the best out of your hardware depends on questions like these, so you have to analyse your Apache logs and tune appropriately. Tuning Apache, mysql and PHP are all subjects in their own right. For FreeBSD, read tuning(7). Are you running FreeBSD ATM? Then some numbers from iostat, top etc. would be useful in analysing how your new server is going to cope and how much spare capacity you'll have, but the numbers are dependent on how you've tuned it (if at all). Hope I've given you something to think a
RAID-5/Stripe Size and verifying absolute disk stripe access
My problem is aligning reads/writes properly on a 3 disk RAID-5 volume with stripe size of 16384. Since my measurements all show the same relatively low read/write performance on the volume matter which offset i choose on the disklabel "partition" (i've tried with the granularity of a single sector and the results are the same). In FreeBSD 7.0-RELEASE, is there any way to find out a absolute read/write sector location for the beginning of filesystem within a disklabel "partition"? -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Updating older 7.0 to newer sources, local buildworld problems
Jason C. Wells wrote: Sten Daniel Soersdal wrote: I'm having troubles updating my buildworld and i've had troubles narrowing it down on google. My question is: Is there any way i could update the system in the state it is now? (Preferably without reinstalling the entire system from scratch.) PS. I've heard about freebsd-update but never got it to work (see below) I have never used freebsd-update so I can't comment there. I think that's a binary update though. You are trying to build from source. I would use cvsup to update my source tree to 7.0-RELEASE. Since your initial sources are before Sep 19, they are not 7.0-RELEASE. Read about cvsup in the handbook. Look at the example files in /usr/share/examples/cvsup. The I would run 'make clean' and 'make buildworld' all over again. Except for creating a newfs and tweaking files in /etc, building the world is pretty much installing the entire system. http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/cutting-edge.html Be careful about the cvs tag that you use. RELENG_7_0_0_RELEASE is what you will eventually want. If that tag has not yet been set by releng@, then you will delete your sources. I recommend '-d 20' with cvsup to protect yourself from accidental deletions. RELENG_7 is probably what you want until 7.0 is actually released. Thank you, i appreciate your time. At the bottom of the (messy) mail i sent there was the error i get from building sources. I think it's the same error, no matter what version of sources i download. Could it have any relations to any compiler version upgrades done in 7.0 (not by me but by freebsd project)? -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Updating older 7.0 to newer sources, local buildworld problems
/src/gnu/usr.bin/cc/cc_tools/../../../../contrib/gcc -I/usr/src/gnu/usr.bin/cc/cc_tools/../../../../contrib/gcc/config -I/usr/src/gnu/usr.bin/cc/cc_tools/../../../../contrib/gcclibs/include -I/usr/src/gnu/usr.bin/cc/cc_tools/../../../../contrib/gcclibs/libcpp/include -I/usr/src/gnu/usr.bin/cc/cc_tools/../../../../contrib/gcclibs/libdecnumber -g -DGENERATOR_FILE -DHAVE_CONFIG_H -I/usr/obj/usr/src/tmp/legacy/usr/include -c /usr/src/gnu/usr.bin/cc/cc_tools/../../../../contrib/gcc/gengtype.c In file included from ./tm.h:4, from /usr/src/gnu/usr.bin/cc/cc_tools/../../../../contrib/gcc/gengtype.c:24: ./options.h:901: error: redeclaration of enumerator 'OPT_w' ./options.h:899: error: previous definition of 'OPT_w' was here *** Error code 1 Stop in /usr/src/gnu/usr.bin/cc/cc_tools. *** Error code 1 Stop in /usr/src. *** Error code 1 Stop in /usr/src. *** Error code 1 Stop in /usr/src. -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: RAID mirror really worked
Wojciech Puchar wrote: gmirror works too very good without any hardware :) CB> Yes, but a hardware RAID works without the OS having to know about it. :-) ...and its failures? ;) :) for mirroring there is almost no CPU overhead so buying extra hardware doesn't make sense at all. not mentioning that most of such hardware are actually normal disk controllers with extra soft in BIOS. these are supported by ataraid driver. much better is to use gmirror so it will be completely portable. and - with gmirror you DO NOT have to mirror/stripe/concat whole drives. and that's what i do most often - mirror important data but store unimportant data without it. ... But with a raid controller mirror you do not have to send the data twice over the host bus. gmirror is awesome -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: how to be *nix programmer
Radheshyam Bhatt wrote: Hello People, How's it going?I am interested in to developing drivers for FreeBSD. How do I go about start learning program for that? What books & resources I should look in to. I know C, and I am learning about processes, and system calls. Also where would I take my questions to if I don't get something and need help for something in system's programming... Please email me back.. I would recommend reading: "The Design and Implementation of the FreeBSD Operating System" by by Marshall Kirk McKusick (Author), George V. Neville-Neil (Author) The questions go where they are appropriate (scsi driver? perhaps freebsd-scsi@ or freebsd-hackers@ ?). -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: (postfix) SPAM filter?
Rudy wrote: Steve Bertrand wrote: * Once it is setup then it would require no additional maintenance. * Potential spam messages are marked with a special header that can be filtered on user discretion on their local mail client software. Yes, one recommendation for sure. Give up on your first goal. It'll never happen, because fighting spam is an arms race, with new tactics needing to be adopted. Amen (or Ahem, or what BSDie would say). There will *ALWAYS* be maintenance. If you are not developing new regexs and/or solutions to fight the daily produced techniques that make up SPAM, then you are implementing them. I have found spam assassin with nightly updates of the helpful (there are other people developing new regexs daily). 48 5 * * * /usr/local/bin/sa-update --channel updates.spamassassin.org && /usr/local/etc/rc.d/sa-spamd restart There are other channels you can subscribe to. Another super helpful bocker is to block all inbound connections from IPs without reverse DNS. Don't forget to virus check your email while you are at it -- there are several packages (clamav is one). And finally, a couple of RBLs added into the mix are helpful. Awesome, i didn't see the subscriptions on their website. This is exactly what i need. -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
(postfix) SPAM filter?
We have a need for a relatively painless anti-spam solution that would reduce the amount of incoming spam (via postfix mail router). The problem is that i have little knowledge on what this actually means. Googling reveals a whole "universe" of interesting ways but what should i pursue? The things that are important to me is: * Once it is setup then it would require no additional maintenance. * Potential spam messages are marked with a special header that can be filtered on user discretion on their local mail client software. Neither performance, scalability, license nor cost is of much importance to me at this point. Any hints? -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Deny access from localhost to internet.....
Agus wrote: Hi guys, How are you today? The question is this..I want to restrict external access, that is from my BSD to the internet, to some groups of users. Other groups i want to access internet normally. I dont want this group of users to be able to establish connections to the internet but yes to the internal systems on the LAN... Is this possible without hacking the kernel? Thanks and salutes for all You want to restrict internet, but not LAN, access for certain users logged into your BSD box? man ipfw( look for "uid" and "gid" ) man pf ( look for "user" and "group" ) -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
PR/51920: Please add Norwegian Collate support to FreeBSD
Today i found a very old but very useful PR which gave my FreeBSD servers the ability to properly sort strings containing characters in the norwegian alphabet. The default (la_LN.ISO8859-1/15) sorts by ASCII value which is not quite right. http://www.freebsd.org/cgi/query-pr.cgi?pr=conf/51920 I had to manually update the Makefile but the .src files where copied as is. I haven't found anything wrong with them, but I'll be happy to supply an update if anyone does. Please take a look at this old issue and put this PR to rest. -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: IPMI A Question to all Dell Users.
Grant Peel wrote: Hi all, I have been reading a bit about IPMI. I am running 6.2 on all my servers. Does any Dell (PowerEdge) users have the IPMI port installed? Is it safe? Easy to use? Any problems with installation? I am mostly interested in viewing sensor info and extracting SELs. TIA, Have been using IPMI on PE2850 without problems. Although i have not performed any extensive testing on the functionality, it appears correct and i have had no trouble with it. I have used /usr/ports/sysutils/ipmitool on 6.1. /usr/ports/sysutils/ipmi-kmod is either available as ports module or in base system on some versions of FreeBSD. -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: IDE ultraDMA problem
Mario Lobo wrote: Hello to all; I had a MSI mobo 645 Ultra with 1.5G ram, pentium 4 1.7 Ghz, 3 IDE HD, 1 SAMSUNG 80 G, 1 SAMSUNG 120 G, maxtor 120 G and a LG DVD writer. FreeBSD 6.2 recognized all HDs as ultraDMA 100. Fine. Then a bought a ASUS p5vd2-x, 1G ram, Gforce 7200 video(pci-e), pentium D 940 Dual core and kept the same drives. After tuning and recompiling the kernel a couple times, I got almost everything working great ! SMP, acpi, network, you name it, EXCEPT ultraDMA. if I leave: hw.ata.atapi_dma: 1 hw.ata.ata_dma: 1 FreeBSD reports them (all) at ultraDMA 33. The system boots ok but after 7 or 10 minutes (even if doing nothing), I start geting messages from g_vfs_(); WRITE DMAERROR that can come from any of the drives until the system becomes unstable and ends up rebooting itself. if I leave: hw.ata.atapi_dma: 0 hw.ata.ata_dma: 0 The HDs get down to PIO 4 and the system works fine but at an incredible performance cost. Finacial issues force me to make this work instead of buying more stuff. Thanks for any suggestions, Sorry, i that should be atacontrol cap not atacontrol info -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: IDE ultraDMA problem
Mario Lobo wrote: Hello to all; I had a MSI mobo 645 Ultra with 1.5G ram, pentium 4 1.7 Ghz, 3 IDE HD, 1 SAMSUNG 80 G, 1 SAMSUNG 120 G, maxtor 120 G and a LG DVD writer. FreeBSD 6.2 recognized all HDs as ultraDMA 100. Fine. Then a bought a ASUS p5vd2-x, 1G ram, Gforce 7200 video(pci-e), pentium D 940 Dual core and kept the same drives. After tuning and recompiling the kernel a couple times, I got almost everything working great ! SMP, acpi, network, you name it, EXCEPT ultraDMA. if I leave: hw.ata.atapi_dma: 1 hw.ata.ata_dma: 1 FreeBSD reports them (all) at ultraDMA 33. The system boots ok but after 7 or 10 minutes (even if doing nothing), I start geting messages from g_vfs_(); WRITE DMAERROR that can come from any of the drives until the system becomes unstable and ends up rebooting itself. if I leave: hw.ata.atapi_dma: 0 hw.ata.ata_dma: 0 The HDs get down to PIO 4 and the system works fine but at an incredible performance cost. Finacial issues force me to make this work instead of buying more stuff. Thanks for any suggestions, There are usually one of two reasons for this; 1) Cable not being correct for slots (which UDMA33 suggests) 2) Controller not supported by driver (the controller does/requires something special) I have had perfectly working cables not work just because the motherboard manufacturer decided they wanted to reorder some of the pins on the motherboard (and i wasn't using cables that came with the motherboard). Perhaps you could post 'pciconf -lv' together with 'atacontrol list' and 'atacontrol info ' where is the names of the devices connected. E.g. acd0 ad0 ad1 etc.. dmesg should probably help? -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
TMPFS, is it available on FreeBSD 6 or 7?
Is TMPFS available on FreeBSD 6 or 7? Or do i have to settle for mfs? [venting frustration] The reason i ask is i need something very flexible when it comes to memory usage as a poor-mans shared-memory feature for a closed source data conversion app, used only occasionally but with *alot* of data passing through. By alot i mean it took our quad core xeon 3 ghz with 4 gbyte memory and 4 SAS drives in 1+0 RAID, about a day and a half to process (on disk). Simple tests (by others) using netbsd on a single cpu 3 ghz and tmpfs, it took roughly 70 minutes. Most of which was spent loading and unloading the data over the network. I'd really like that memory back after use for other things as this server is also serving databases with files etc. (PS. If anyone knows of some place in the world it is legally permissible to shoot a programmer for incompetence/excessive-tmp-usage, then please let me know, that's where are going on company vacation) [/venting] -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: 10Mbps versus 100Mbps Cable Modems
fbsd2 wrote: Comclark cable in Angeles City Philippines has changed from using 100Mbps Cable Modem to 10Mbps Cable Modem. To me this seems to be all wrong as all I see is slower response. Is there any technical or performance reason for any cable internet provider to downgrade their network subscribers cable modems from 100Mbps to 10Mbps? That reason could be compatibility. If you see slower response then perhaps something is wrong. Perhaps you should call their support and verify that you do not have a mismatched duplex setting? Mismatched duplex can come from misbehaving autonegotiation or that one end is set to full-duplex while the other end is set to half-duplex, or, one end is set to full-duplex and the other end is set to auto-negotiate (which results in falling back to half-duplex). -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: pf(4) + fetch(1) + http://ftp.gnu.org
Vlad GURDIGA wrote: Hello, There is one strange thing going on with this combination. I saw this many times by now: when fetch(1) is trying to download something from http://ftp.gnu.org, it is hanging after a very small amount of data; sometimes on 0%. After disabling pf(4), fetch(1) is not hanging any more, so I guess that the problem is somewhere in my pf.conf. Here is it: pf.conf -- begin --- ext_if = "em0" icmp_types="echoreq" # don't filter on the loopback interface set skip on lo0 set block-policy return scrub all no-df random-id reassemble tcp # setup a default deny policy block all # activate spoofing protection for the internal interface. antispoof quick for lo0 inet # pass tcp, udp, and icmp out on the external (Internet) interface. # keep state on udp and icmp and modulate state on tcp. pass in on $ext_if proto tcp from any to $ext_if port 65522 keep state pass in inet proto icmp all icmp-type $icmp_types keep state pass out on $ext_if proto tcp all modulate state flags S/SA pass out on $ext_if proto {udp, icmp} all keep state pf.conf -- end --- My guess is, your path-mtu is lower at some point between you and ftp.gnu.org which (may or may not) result in an icmp packet indicating that the packet was dropped as fragmentation was needed but DF flag was set. it is usually possible to see via tcpdump or ethereal. try permitting all icmp packets both ways to see. -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: detect stderr writes when combined
Robin Becker wrote: This is possibly a stupid question, but I would like to have a particular sh script stdout and stderr combined, but at least detect when stderr has been used. In particular for my cron scripts it seems that error messages get wrapped up and emailed, but they are then out of context with the normal informative output. If I could have the combined output go to my logs in the normal way and detect that some error output had occurred I could then email the whole cron output. I'm fairly sure this is doable with some C, but is there an easier way? Perhaps using redirects? this can be used in /bin/sh (and thus cron): ./myprogram 2>&1 where "myprogram" is your application. this would combine stderr and stdout. -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Reading temperature on DELL PE2850?
I'm contemplating setting "sysctl machdep.cpu_idle_hlt=0" on my 4 core PE2850, but i am concerned that without HLT-ing would increase the temperature levels to dangerous levels. Are there any ways for me to read temperatures from FreeBSD/DRAC4 ? -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: IRQ storm
Vlad GURDIGA wrote: > On 27/03/07, Sten Daniel Soersdal <[EMAIL PROTECTED]> wrote: >> Vlad GURDIGA wrote: >> > After running this: >> > >> > /sbin/atacontrol reinit ata2 >> > >> > the storm was gone. >> > My HDD in on ata4: >> > >> > #atacontrol info ata4 >> > Master: ad8 Serial ATA II >> > Slave: no device present >> > >> > >> > Why? >> > >> >> Could be one of any number of reasons and my guess is that the drive >> wants to enter sleep mode, due to overheating or energy saving. And >> either controller, driver or both found it unreasonable. > > You can see the brand and model of my drive. I have Windows XP on the > same computer which works fine. Would the drive want to enter sleep > mode only under FreeBSD? Depends on your workload. Perhaps you do something under FreeBSD that generates more heat? >> Perhaps you could inform us after you have researched it? > > I do not have a plain idea about what I could or should research in > this case... Could you please be morespecific? I would start with the name of the SATA controller and drive and see whether someone else has had the same problems (search archives and google). Then i'd probably check temperatures, chipsets, try -CURRENT (for driver updates to see whether that fixes it) and go from there. See if you can identify a condition that triggers the IRQ storm. -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: IRQ storm
Vlad GURDIGA wrote: > After running this: > > /sbin/atacontrol reinit ata2 > > the storm was gone. > My HDD in on ata4: > > #atacontrol info ata4 > Master: ad8 Serial ATA II > Slave: no device present > > > Why? > Could be one of any number of reasons and my guess is that the drive wants to enter sleep mode, due to overheating or energy saving. And either controller, driver or both found it unreasonable. Perhaps you could inform us after you have researched it? -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Best partitioning scheme for my HDD? Please advise.
Joe Vender wrote: > I have a 6120MB HDD which will be dedicated to FreeBSD 6.2. I intend to > install the ports collection and also KDE. I will operate from the KDE > environment using FreeBSD as a standalone desktop machine connected to the > net via a dialup internet connection. What would be the best sizes for the > disk partitions so that I don't run out of space on any of them while also > leaving the maximum amount of space possible for the future software to be > installed? > My partitions will be: > / > swap > /var > /tmp > /usr > > as suggested using the auto option during slice creation. > > I've found that if I use the default sizes that are chosen by the installer > using the auto option, the /usr partition fills up before everything is > installed and the installation fails. If I remember correctly, the auto > feature sets the sizes around the following sizes for my HDD: > / ~500MB > swap ~600MB > /var ~1300MB > /tmp ~ 500MB > /usr ~3GB > Perhaps it is possible to reduce / to ~128 mb? boot up and see estimate that you need several mb free and estimate double of /boot/kernel directory (if you install new kernel you get to keep the old one, therefore estimate double that directory size). Swap ... well for a successful crash dump you need as much as swap as you have memory. /var can be further reduced to perhaps 512mb or perhaps even less? if you do buildworld then /var/tmp/ might be used more. Perhaps you could merge / and /var and symlink /var/tmp with say /usr/tmp ? /tmp can be reduced, you can even remove it completely and use tmpfs for that partition. /usr needs the rest Just a (cluttered) suggestion. The cool part about FreeBSD is what you can get away with, just by symlinking (ln -s) a few folders here and there. After all, it's just your desktop, right? -- Sten Daniel Soersdal ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"