Re: soft update should be default
On Fri, May 04, 2001 at 10:39:51PM -0400, Chris BeHanna wrote: On Fri, 4 May 2001, Tadayuki OKADA wrote: [...] I've heard that it always keeps consistency. So you can skip fsck after the crash. #I don't know the detail, so please someone correct me if I'm wrong. I've had a number of crashes recently while trying to get my new Thunderbird box up and running with the disk from my old box. I have softupdates enabled on two partitions, yet they still fsck on the way back up after a crash (and, at 18GB each, it takes awhile). Next time I play games, I'm mounting them read-only first! Boot single-user, mount -f. You will still want to fsck, some day. I have a server with two .5TB partitions. I make their fstab entries noauto so if the machine does crash, it comes back up, and I can mount -f or fsck at my discretion. -danny To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
What's the conclusion? (was soft update should be default)
On Fri, 04 May 2001 18:42:54 -0700 Mike Smith [EMAIL PROTECTED] wrote: Why 'soft update' is not default? It adds performance and stability, doesn't it? It requires disabling of write caching, which typically reduces performance (significantly). If this is the only problem, I think softupdates should be the default. Because: 'write caching' is not the default. Anyone who wants it can disable softupdates. Besides it seems 'write caching' is not recommended for the usual usage. If you have other resons not to do so, could you please explain it? -- Tadayuki OKADA To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Tagged Queueing and ATA driver (was soft update should be default)
Gee, maybe in another few years IDE will have implement the *entire* SCSI command set! Now wouldn't that be progress! Not! You will never see me turn on tagged queueing for IDE. If performance is an issue, SCSI is the solution. I want my data cooked over-easy thank you very much, not scrambled! -Matt To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: soft update should be default
[EMAIL PROTECTED] said: WCE very dangerous. WCE very dangerous. WCE very dangerous. True, but since the new ATA driver was installed in STABLE about February 25, WCE has also become mandatory for my NEC Versa 6050 MX laptop. If I don't set hw.ata.wc=1, kernel buffers become corrupted. Perhaps the previous driver was silently setting write caching too, but things just don't feel as secure anymore. -- M/S 258-5 | 1024-bit PGP fingerprint: | [EMAIL PROTECTED] NASA Ames Research Center | 41 B0 89 0A 8F 94 6C 59 | (650) 604-4416 Moffett Field, CA 94035-1000 | 7C 80 10 20 25 C7 2F E6 | FAX: (650) 604-4377 We each earn what freedom of speech we defend for those who most offend us. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re[2]: soft update should be default
On Sun, 6 May 2001, Gabriel Ambuehl wrote: -BEGIN PGP SIGNED MESSAGE- Hello Charlie, Sunday, May 06, 2001, 1:53:20 AM, you wrote: I see the same behaviour on one of those disks, too. But - aren't IBM's DTLA-series disks the only IDE drives that support TCQ? [ It's a -very- SCSI-feeling feature, in my mind. ] It was about time this was implemented in IDE disk as this takes them nearer to the ridiculously overpriced SCSI stuff. While I've always been a lover of IBM drives, this feature is really cool. Is it on by default if you use DTLA disks or do I need to activate it specifically, BTW? In /boot/loader.conf: hw.ata.tags=1 I have several other random stupid IDE drives in this machine, too, and don't notice any adverse behaviour. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re[2]: soft update should be default
On Sun, 6 May 2001, Gabriel Ambuehl wrote: -BEGIN PGP SIGNED MESSAGE- Hello Charlie, Sunday, May 06, 2001, 1:53:20 AM, you wrote: IBM DLTA-307030 Ultra ATA drive (tags/no WC vs. no tags/WC). With neither option, it is terrible, of course. I see the same behaviour on one of those disks, too. But - aren't IBM's DTLA-series disks the only IDE drives that support TCQ? [ It's a -very- SCSI-feeling feature, in my mind. ] It was about time this was implemented in IDE disk as this takes them nearer to the ridiculously overpriced SCSI stuff. While I've always been a lover of IBM drives, this feature is really cool. Is it on by default if you use DTLA disks or do I need to activate it specifically, BTW? Of course, I should read all of my E-mail before responding. From other messages on freebsd-stable today: TCQ also turns on a -form- of write caching, it may not be quite the same as normal, but it still does, and should probably be considered not-totally safe. Bah. Sorry. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: soft update should be default
Nick Barnes wrote: This sounds as if there isn't _any_ way for the kernel (or, better, an application) to make sure that its bits have got written. Is that really true? Shouldn't the man pages for fsync(1), fsync(2), and sync(8) reflect this? sync(2) has something under BUGS Sure there is. 1. Disable write-caching on your drive. Currently this is the default anyway. There is a sysctl to control it. 2. Turn off soft-updates. 3. At that point, so far as I know, sync should work correctly. If this is true, it's not good. Presumably fsync(2) will get the data down the cable to the disk unit. If the CPU, kernel, etc goes toes-up a microsecond later, will my bits still hit the platter? They will if write caching in the disk is turned off AND the write operation actually completed (that is, the drive acknowledged completion). I'm assuming I can keep the power on, which is a separate and well-understood problem. But if there's a panic and reboot, presumably there's some kind of reset now message sent to the disk unit (the exact details no doubt depend on the disk type). I believe the disk write caches are only in danger if power is unexpectedly lost. If you actually halt the OS, the disk will have time to finish the writes. Will it write my bits or flush them? How do different disk units compare in this respect? When I started with FreeBSD, the general understanding was that people who cared about data integrity used SCSI, people who really cared used RAID on SCSI, and people who were fanatical about it used hardware SCSI-to-SCSI RAID in a separate rack unit with redundant PSUs and controllers and very high-quality cables. Is this still the received wisdom? I think with write-caching turned off and softupdates anyone should be happy unless they really require transactional recording (that is, if a transaction is acknowledged, then it must not ever be lost). Those folks should probably be running Oracle on a raw partition or some such. :-) With softupdates and no write-cache, I was able to start an untar of the ports tree (truly one of the worst-case scenarios -- lots of directories and tiny files), press RESET and come back up without much hassle. The net result of the fsck was that the filesystem was recovered to an exact moment in time slightly before when I actually hit RESET. That is, the metadata cache lagged behind realtime slightly, but was maintained perfectly. The cutoff between the preserved state and the lost state was perfectly temporal. That's why I would suggest those interested in transaction assurance not use softupdates. :-) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: soft update should be default
Date: Sat, 05 May 2001 09:31:09 -0700 From: Nick Sayer [EMAIL PROTECTED] Sender: [EMAIL PROTECTED] That may be the original intent, but cheap IDE drives let you turn on write caching, and they're for sure not battery-backed (nor do they attempt to store enough power at power-off to write back the cache with the remaining rotational latency or any such trickery). They lie about it. Write caching is evil unless you specifically know that it's being battery backed. 99.44% of the time, that's not the case. An obvious exception is the laptop. I always turn on write cache on my laptop as I know that it has a LONG battery backup. For a worst-case type of operation (dd), I get 4x faster writes with write-cache enabled on my laptop. R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: [EMAIL PROTECTED] Phone: +1 510 486-8634 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message