Re: fstab weirdness / UPDATING
Yup, seems like i goofed that patch. Can you try this for me ? Index: preen.c === RCS file: /home/ncvs/src/sbin/fsck/preen.c,v retrieving revision 1.22 diff -u -r1.22 preen.c --- preen.c 2001/03/30 08:01:34 1.22 +++ preen.c 2001/04/07 07:01:47 @@ -178,7 +178,6 @@ continue; } - if (WIFEXITED(status)) retcode = WEXITSTATUS(status); else @@ -215,6 +214,9 @@ if (TAILQ_EMPTY(d-d_part)) { TAILQ_REMOVE(diskh, d, d_entries); ndisks--; + } else { + if ((ret = startdisk(d, checkit)) != 0) + return ret; } } if (flags CHECK_DEBUG) { -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: fstab weirdness / UPDATING
Date: Sat, 07 Apr 2001 09:08:41 +0200 From: Poul-Henning Kamp [EMAIL PROTECTED] Yup, seems like i goofed that patch. Can you try this for me ? OK; it works much better with the patch: it seems to work correctly, from what I've tested so far. [Below is a sketch of what I did, so folks will be able to judge how much trust to place in the above assessment.] * Booted -CURRENT into single-user mode. Even though this was a clean reboot, I did a (complete) "fsck" before "mount -a", then fired up tcsh to use as the shell for the following. * I applied the patch from within the /urs/src/sbin/fsck directory. (Yeah, I used -C first to make sure patch wouldn't whine too much.) * Since the patch only affects the logic of a user-level program (vs. the kernel, for example), I went ahead typed "make" while I was there. * cd /sbin mv fsck{,.save} mv /usr/obj/usr/src/sbin/fsck/fsck . ls -l fsck* chmod u-w fsck * Re-boot -CURRENT, single-user mode. "fsck -p" now claims to check each filesystem. Good. * mount -a; things look good. Power-cycle the machine. (This is still single-user mode, so the filesystems should merely have been marked dirty because they're mounted, but nothing should actually have been doing anything to them other than that.) * Re-boot -CURRENT, single-user mode. "fsck-p" again does the usual and expected fsck behavior -- all filesystems are checked flagged "clean". I like it. :-) Thanks, david -- David H. Wolfskill [EMAIL PROTECTED] As a computing professional, I believe it would be unethical for me to advise, recommend, or support the use (save possibly for personal amusement) of any product that is or depends on any Microsoft product. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: fstab weirdness / UPDATING
Hello everybody, I am sorry to report that the fix committed to preen.c by phk recently did not fix the problem for me entirely. I have: ad0 with partitions within s1 a through h (ad0s1a-ad0s1h) and ad1s1a. Upon rebooting (after a clean shutdown) with an up-to-date kernel and userland, all the partitons on ad0 were checked, but the ad1 was not. Yet, it was mounted r/w afterwards. I did not yet test in single-user mode though what if I say 'fsck -p'... -- Regards: Szilveszter ADAM Szeged University Szeged Hungary To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: fstab weirdness / UPDATING
Hello, On Sat, Apr 07, 2001 at 12:04:07PM -0700, David Wolfskill wrote: I suggest "fsck -d -p" -- the debugging output could prove useful. Here we go: pass1 pass1, name /dev/ad0s1a pass2 pass2, name /dev/ad0s1f pass2, name /dev/ad0s1g pass2, name /dev/ad0s1h pass2, name /dev/ad0s1d pass2, name /dev/ad0s1e Parallel start disk /dev/ad0: ad0s1f, ad0s1g, ad0s1h, ad0s1d, ad0s1e start /tmp nowait fsck_ufs -p /dev/ad0s1f Parallel wait done ufs: /dev/ad0s1f (/tmp) 0x0 start /usr nowait fsck_ufs -p /dev/ad0s1g Parallel wait done ufs: /dev/ad0s1g (/usr) 0x0 start /usr/local nowait fsck_ufs -p /dev/ad0s1h Parallel wait done ufs: /dev/ad0s1h (/usr/local) 0x0 start /var nowait fsck_ufs -p /dev/ad0s1d Parallel wait done ufs: /dev/ad0s1d (/var) 0x0 start /var/log nowait fsck_ufs -p /dev/ad0s1e Parallel wait done ufs: /dev/ad0s1e (/var/log) 0x0 Parallel end In other words, nothing unusual, apart from the fact that ad1 is not even mentioned. If I after this say 'fsck -d -p /dev/ad1s1a', it says: start /dev/ad1s1a wait fsck_ufs -p /dev/ad1s1a and then exits, seemingly succesfully. Nothin more... if there is anything I can be helpful with, please tell me. -- Regards: Szilveszter ADAM Szeged University Szeged Hungary To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: fstab weirdness / UPDATING
In message [EMAIL PROTECTED], Szilveszter Adam writ es: Hello, On Sat, Apr 07, 2001 at 12:04:07PM -0700, David Wolfskill wrote: I suggest "fsck -d -p" -- the debugging output could prove useful. Here we go: Is this with or without the patch I committed earlier today ? -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: fstab weirdness / UPDATING *solved!*
Hello guys, [Cc-ing the list just for the sake of the archives, so that people could see what the resolution of the case was.] Thanks for trying to help me while I was busy sleeping... here I am again. As suggested, I looked at my fstab, although I do not remember fiddling with it in a long time. (In fact, the last time was when I installed my second disk into the system some time at the end of last year.) And sure as hell, there it was: next to /dev/ad1s1a there was *no* dump number and *no* pass no. Not zeros, but nothing, nada, nichts. This, of course, explains a lot of things... after putting the correct values in there, fsck worked. But I have no idea how/why the numbers got deleted... and then why it worked before? Thanks all the same, things seem to work out A OK right now. :-) -- Regards: Szilveszter ADAM Szeged University Szeged Hungary To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: fstab weirdness / UPDATING
OK; I finally got a few consecutive minutes to look over fsck as of today's -CURRENT. In replying to different message, I had earlier reported what some others noted a couple of days ago: that not all of the filesystems are checked at boot (only 2 are), and experimentation demonstrated that in "preen" mode (-p), this is what happens, but it is peculiar to preen mode: bare "fsck" (with no arguments) does check all filesystems. First, for reference, here's the fstab in question (yes, it's fairly chopped up): # DeviceMountpoint FStype Options DumpPass# /dev/ad0s3b noneswapsw 0 0 /dev/ad0s1a /S1 ufs rw 2 2 /dev/ad0s1e /S1/usr ufs rw 2 2 /dev/ad0s2a /S2 ufs rw 2 2 /dev/ad0s2e /S2/usr ufs rw 2 2 /dev/ad0s3a / ufs rw 1 1 /dev/ad0s3e /usrufs rw 2 2 /dev/ad0s3g /varufs rw 2 2 /dev/ad0s3h /common ufs rw 2 2 /dev/acd0c /cdrom cd9660 ro,noauto 0 0 proc/proc procfs rw 0 0 So I booted -CURRENT into single-user mode (after a clean reboot); here's a hand-generated transcription: Enter full pathname of shell or RETURN for /bin/sh: # fsck -d -p pass 1 pass 1, name /dev/ad0s3a start / wait fsck_ufs -p /dev/ad0s3a pass 2 pass 2, name /dev/ad0s1a pass 2, name /dev/ad0s1e pass 2, name /dev/ad0s2a pass 2, name /dev/ad0s2e pass 2, name /dev/ad0s3e pass 2, name /dev/ad0s3g pass 2, name /dev/ad0s3h Parallel start disk /dev/ad0: /dev/ad0s1a /dev/ad0s1e /dev/ad0s2a /dev/ad0s2e /dev/ad0s3e /dev/ad0s3g /dev/ad0s3h start /S1 nowait fsck_ufs -p /dev/ad0s1a Parallel wait done ufs: /dev/ad0s1a (/S1) = 0x0 Parallel end disk /dev/ad0: /dev/ad0s1e /dev/ad0s2a /dev/ad0s2e /dev/ad0s3e /dev/ad0s3g /dev/ad0s3h # I hope this is useful. I should be able to perform additional testing. Thanks, david -- David H. Wolfskill [EMAIL PROTECTED] As a computing professional, I believe it would be unethical for me to advise, recommend, or support the use (save possibly for personal amusement) of any product that is or depends on any Microsoft product. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: fstab weirdness / UPDATING
Hello, On Wed, Apr 04, 2001 at 10:49:22PM -0700, Alfred Perlstein wrote: I've had that problem on one box also -- when the boot occurs, my /var partition is not fsck'd. It's the second pass-two file system, which means that it *should* be checked :-). I suspect a nit in the recent fsck cleanup, so I'm CC'ing phk, whose mailbox is obviously too empty. I think I'm seeing the same thins... At boot I see a kernel printf "warning: /var was not properly dismounted" then /var mounts. If I unmount it and fsck it it's dirty. I saw something similar too, but did not get around to posting yet because I wanted to understand what's going on. On boot, I see only two of my partitions being reported as clean, and not the others. After this, all of them are mounted allright. (They happen to be ad0s1a and ad0s1f don't know why...) I don't know if it causes filesystem corruption, because it is my policy to always 'boot -s' upon unclean shutdown and fsck all partitions manually. (In fact I just did it again because X has a tendency to freeze the machine off and on for unknown reasons... maybe related?) If this is an fsck problem, then I figured there was no big problem if your previous shutdown was clean... am I right? Or maybe there is a problem with shutdown too that fsck simply doesn't notice? Brrr...:-) -- Regards: Szilveszter ADAM Szeged University Szeged Hungary To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
fstab weirdness / UPDATING
My understanding is that there have been some changes regarding fstab, fsck, etc. recently. On boot, I find that only first two filesystems are listed as "clean", and on a recent un-clean shutdown only the first two were properly fsck'd (although the system was allowed to boot and mount read-write the unclean filesystems). I can only assume that some change to the boot procedure was made, since I made no changes. Are we going to see an UPDATING entry regarding this? I can't seem to find any explicit instructions in the mailing lists either. It worries me a bit when unclean filesystems are being mounted rw. WNM -- _ __ ___ ___ ___ ___ Wesley N Morgan _ __ ___ | _ ) __| \ [EMAIL PROTECTED] _ __ | _ \._ \ |) | FreeBSD: The Power To Serve _ |___/___/___/ 6bone: 3ffe:1ce3:7::b4ff:fe53:c297 Hi! I'm a .signature virus! Copy me into your ~/.signature to help me spread! To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: fstab weirdness / UPDATING
I've had that problem on one box also -- when the boot occurs, my /var partition is not fsck'd. It's the second pass-two file system, which means that it *should* be checked :-). I suspect a nit in the recent fsck cleanup, so I'm CC'ing phk, whose mailbox is obviously too empty. Robert N M Watson FreeBSD Core Team, TrustedBSD Project [EMAIL PROTECTED] NAI Labs, Safeport Network Services On Wed, 4 Apr 2001, Wesley Morgan wrote: My understanding is that there have been some changes regarding fstab, fsck, etc. recently. On boot, I find that only first two filesystems are listed as "clean", and on a recent un-clean shutdown only the first two were properly fsck'd (although the system was allowed to boot and mount read-write the unclean filesystems). I can only assume that some change to the boot procedure was made, since I made no changes. Are we going to see an UPDATING entry regarding this? I can't seem to find any explicit instructions in the mailing lists either. It worries me a bit when unclean filesystems are being mounted rw. WNM -- _ __ ___ ___ ___ ___ Wesley N Morgan _ __ ___ | _ ) __| \ [EMAIL PROTECTED] _ __ | _ \._ \ |) | FreeBSD: The Power To Serve _ |___/___/___/ 6bone: 3ffe:1ce3:7::b4ff:fe53:c297 Hi! I'm a .signature virus! Copy me into your ~/.signature to help me spread! To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: fstab weirdness / UPDATING
* Robert Watson [EMAIL PROTECTED] [010404 22:04] wrote: I've had that problem on one box also -- when the boot occurs, my /var partition is not fsck'd. It's the second pass-two file system, which means that it *should* be checked :-). I suspect a nit in the recent fsck cleanup, so I'm CC'ing phk, whose mailbox is obviously too empty. I think I'm seeing the same thins... At boot I see a kernel printf "warning: /var was not properly dismounted" then /var mounts. If I unmount it and fsck it it's dirty. -- -Alfred Perlstein - [[EMAIL PROTECTED]|[EMAIL PROTECTED]] Represent yourself, show up at BABUG http://www.babug.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message