[zfs-discuss] unallocated block pointers in indirect block
Guys, I am curious if the following is something valid or a result of a bad corruption. Here is a snippet from a dump of an indirect block (zdb -R with i flag): ... [many uninterestingly good BPs skipped] DVA[0]=0:13da6a47a00:2d000 [L0 ZFS plain file] fletcher4 uncompressed LE contiguous unique single size=2L/2P birth=39369L/39369P fill=1 cksum=3df551808443:f842dcdc8a32e57:af4094c9cac4733c:c7ac3865df0c4ced DVA[0]=0:13da6a74a00:2d000 [L0 ZFS plain file] fletcher4 uncompressed LE contiguous unique single size=2L/2P birth=39369L/39369P fill=1 cksum=3e69cad45d5c:f8bc987d5aabad0:e5f72ba62f3b2955:84a2a388440c0d31 hole hole hole hole hole hole hole hole hole hole DVA[0]=0:ac00:1c400 DVA[1]=0:1ee00:1ea00 DVA[2]=0:1ee00:11400 [L0 unallocated] inherit inherit BE contiguous unique triple size=1c000L/200P birth=12L/71P fill=93 cksum=9a:97:8f:b ... From here on all the BPs appear either as holes or as L0 unallocated. I wonder how those unallocated BPs came to be and if their presence is valid and is to be expected by the ZFS code. Just in case, the indirect block passes checksum verification (and also decompression, in fact). Thank you very much in advance. -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] zdb leaks checking
Does zdb leak checking mechanism also check for the opposite situation? That is, used/referenced blocks being in free regions of space maps. Thank you. -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] cross platform (freebsd) zfs pool replication
on 01/07/2011 00:12 Joeri Vanthienen said the following: Hi, I have two servers running: freebsd with a zpool v28 and a nexenta (opensolaris b134) running zpool v26. Replication (with zfs send/receive) from the nexenta box to the freebsd works fine, but I have a problem accessing my replicated volume. When I'm typing and autocomplete with tab key the command cd /remotepool/us (for /remotepool/users) I get a panic. check the panic @ http://www.boeri.be/panic.jpg Since this is a FreeBSD panic, I suggest that you try getting help on FreeBSD mailing lists, f...@freebsd.org looks like the best choice. BTW, your report doesn't contain your actual panic message and that could be important. - autocomplete tab does a ls command in the background, I think I think there is a problem with NFSv4 acl/id mapping. Normal zfs (inititally created on the FreeBSD box) file systems are working fine. The nexenta box is Active Directory integrated with and the mappings for the users on this cifs share have been created on the fly (Ephemeral ID Mapping). Any solution for this? I really need the ACL permissions to be replicated. So rsync is not a solution. Please help :) root@ ~]# zfs get all remotepool/users NAME PROPERTY VALUESOURCE remotepool/users type filesystem - remotepool/users creation Wed Jun 29 14:42 2011- remotepool/users used 9.06G- remotepool/users available 187G - remotepool/users referenced9.06G- remotepool/users compressratio 1.00x- remotepool/users mounted yes - remotepool/users quota none default remotepool/users reservation none default remotepool/users recordsize128K default remotepool/users mountpoint/remotepool/usersdefault remotepool/users sharenfs off default remotepool/users checksum on default remotepool/users compression off default remotepool/users atime on default remotepool/users devices on default remotepool/users exec on default remotepool/users setuidon default remotepool/users readonly off default remotepool/users jailedoff default remotepool/users snapdir hidden received remotepool/users aclinheritpassthrough received remotepool/users canmount on default remotepool/users xattr off temporary remotepool/users copies1default remotepool/users version 5- remotepool/users utf8only off - remotepool/users normalization none - remotepool/users casesensitivity insensitive - remotepool/users vscan off default remotepool/users nbmandon received remotepool/users sharesmb name=users,guestok=true received remotepool/users refquota none default remotepool/users refreservationnone default remotepool/users primarycache all default remotepool/users secondarycacheall default remotepool/users usedbysnapshots 0- remotepool/users usedbydataset 9.06G- remotepool/users usedbychildren0- remotepool/users usedbyrefreservation 0- remotepool/users logbias latency default remotepool/users dedup off default remotepool/users mlslabel - remotepool/users sync standard default -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] ARC/VM question
I have a semi-theoretical question about the following code in arc.c, arc_reclaim_needed() function: /* * take 'desfree' extra pages, so we reclaim sooner, rather than later */ extra = desfree; /* * check that we're out of range of the pageout scanner. It starts to * schedule paging if freemem is less than lotsfree and needfree. * lotsfree is the high-water mark for pageout, and needfree is the * number of needed free pages. We add extra pages here to make sure * the scanner doesn't start up while we're freeing memory. */ if (freemem lotsfree + needfree + extra) return (1); If I understand correctly, there is no pageout (page daemon is idle) until freemem goes below lotsfree + needfree. So, let's suppose that there is a slow eating userland application such that each time that page scheduler and ARC reclaim thread check freemem its value is in the range from lotsfree + needfree to lotsfree + needfree + extra. This would mean that there would be no paging, but ARC reclaim thread would reduce ARC size and free up some pages. Then the application would eat into those pages again and so on, until ARC size would reach its minimum value. I think that such an outcome would be undesirable, because in my opinion all memory consumers should cooperate in freeing / paging out memory when memory shortage arises. So, do I am miss anything in ARC code or OpenSolaris VM code that would prevent the described scenario from taking place? If not, then why 'extra' was added to the condition in question? I really would like to understand interaction between VM behavior and ARC sizing in OpenSolaris. Thank you very much for any help! -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Encryption?
on 11/07/2010 14:21 Roy Sigurd Karlsbakk said the following: I'm planning on running FreeBSD in VirtualBox (with a Linux host) and giving it raw disk access to four drives, which I plan to configure as a raidz2 volume. Wouldn't it be better or just as good to use fuse-zfs for such a configuration? I/O from VirtualBox isn't really very good, but then, I haven't tested the linux/fbsd configuration... Hmm, an unexpected question IMHO - wouldn't it better to just install FreeBSD on the hardware? :-) If an original poster is using Linux as a host OS, then probably he has some very good reason to do that. But performance and etc -wise, directly using FreeBSD, of course, should win over fuse-zfs. Right? [Installing and maintaining one OS instead of two is the first thing that comes to mind] -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] gang blocks at will?
on 27/05/2010 07:11 Jeff Bonwick said the following: You can set metaslab_gang_bang to (say) 8k to force lots of gang block allocations. Bill, Jeff, thanks a lot! This helped to reproduce the issue and find the bug. Just in case: http://www.freebsd.org/cgi/query-pr.cgi?pr=bin/144214 On May 25, 2010, at 11:42 PM, Andriy Gapon wrote: I am working on improving some ZFS-related bits in FreeBSD boot chain. At the moment it seems that the things work mostly fine except for a case where the boot code needs to read gang blocks. We have some reports from users about failures, but unfortunately their pools are not available for testing anymore and I can not reproduce the issue at will. I am sure that (Open)Solaris GRUB version has been properly tested, including the above environment. Could you please help me with ideas how to create a pool/filesystem/file that would have gang-blocks with high probability? Perhaps, there are some pre-made test pool images available? Or some specialized tool? Thanks a lot! -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] gang blocks at will?
I am working on improving some ZFS-related bits in FreeBSD boot chain. At the moment it seems that the things work mostly fine except for a case where the boot code needs to read gang blocks. We have some reports from users about failures, but unfortunately their pools are not available for testing anymore and I can not reproduce the issue at will. I am sure that (Open)Solaris GRUB version has been properly tested, including the above environment. Could you please help me with ideas how to create a pool/filesystem/file that would have gang-blocks with high probability? Perhaps, there are some pre-made test pool images available? Or some specialized tool? Thanks a lot! -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Oracle to no longer support ZFS on OpenSolaris?
on 23/04/2010 04:22 BM said the following: On Tue, Apr 20, 2010 at 2:18 PM, Ken Gunderson kgund...@teamcool.net wrote: Greetings All: Granted there has been much fear, uncertainty, and doubt following Oracle's take over of Sun, but I ran across this on a FreeBSD mailing list post dated 4/20/2010 ...Seems that Oracle won't offer support for ZFS on opensolaris Link here to full post here: http://lists.freebsd.org/pipermail/freebsd-questions/2010-April/215269.html I am not surprised it comes from FreeBSD mail list. :) Why this attitude about FreeBSD? Did we eat your lunch? Have you actually bothered to follow the link? First, it was a pure speculation inside a question. Second, look what kind of mailing list was that (general FreeBSD-related questions from anyone). Third, look from whom that came - just a random person asking a question. [Paranoia mode: maybe it was even you.] If, for example, I posted some nonsense about e.g. Apple on e.g. OpenSolaris mailing list; what conclusions would you make then? I am amazed of their BSD conferences when they presenting all this *BSD stuff using Apple Macs (they claim it is a FreeBSD, just very bad version of it), Ubuntu Linux (not yet BSD) or GNU/Microsoft Windows (oh, everybody does that sin, right?) with a PowerPoint running on it (sure, who wants ugly OpenOffice if there no brain enough to use LaTeX). What you wrote tells more about you than about FreeBSD and FreeBSD community. P.S. I am surprised that on this useful mostly technical mailing list such random garbage from a random source gets posted at all. And then gets taken seriously even... -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Accidentally added disk instead of attaching
on 06/12/2009 19:40 Volker A. Brandt said the following: I wanted to add a disk to the tank pool to create a mirror. I accidentally used zpool add ? instead of zpool attach ? and now the disk is added. Is there a way to remove the disk without loosing data? Been there, done that -- at a customer site while showing off ZFS. :-) Currently, you cannot remove a simple device. Depending on your Solaris version, you can remove things like hot spares and cache devices, but not simple vdevs. Backup the pool and recreate it in the correct way. Sorry for the bad news -- Volker Yep. My 2 cents -- 'add' and 'attach' are so similar the words that I think that ZFS tools UI designers (if any) should re-consider naming of these commands. Or 'add' command should always be interactive and ask for at least two confirmations that a user knows what he is doing and why. Perhaps, it should include a ZFS micro-exam too. Jokes aside, this is too easy to make a mistake with the consequences that are too hard to correct. Anyone disagrees? -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] feature proposal
What do you think about the following feature? Subdirectory is automatically a new filesystem property - an administrator turns on this magic property of a filesystem, after that every mkdir *in the root* of that filesystem creates a new filesystem. The new filesystems have default/inherited properties except for the magic property which is off. Right now I see this as being mostly useful for /home. Main benefit in this case is that various user administration tools can work unmodified and do the right thing when an administrator wants a policy of a separate fs per user But I am sure that there could be other interesting uses for this. -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] feature proposal
on 29/07/2009 17:24 Andre van Eyssen said the following: On Wed, 29 Jul 2009, Andriy Gapon wrote: Subdirectory is automatically a new filesystem property - an administrator turns on this magic property of a filesystem, after that every mkdir *in the root* of that filesystem creates a new filesystem. The new filesystems have default/inherited properties except for the magic property which is off. Right now I see this as being mostly useful for /home. Main benefit in this case is that various user administration tools can work unmodified and do the right thing when an administrator wants a policy of a separate fs per user But I am sure that there could be other interesting uses for this. It's a nice idea, but zfs filesystems consume memory and have overhead. This would make it trivial for a non-root user (assuming they have permissions) to crush the host under the weight of .. mkdir. Well, I specifically stated that this property should not be recursive, i.e. it should work only in a root of a filesystem. When setting this property on a filesystem an administrator should carefully set permissions to make sure that only trusted entities can create directories there. 'rmdir' question requires some thinking, my first reaction is it should do zfs destroy... -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] feature proposal
on 29/07/2009 17:52 Andre van Eyssen said the following: On Wed, 29 Jul 2009, Andriy Gapon wrote: Well, I specifically stated that this property should not be recursive, i.e. it should work only in a root of a filesystem. When setting this property on a filesystem an administrator should carefully set permissions to make sure that only trusted entities can create directories there. Even limited to the root of a filesystem, it still gives a user the ability to consume resources rapidly. While I appreciate the fact that it would be restricted by permissions, I can think of a number of usage cases where it could suddenly tank a host. One use that might pop up, for example, would be cache spools - which often contain *many* directories. One runaway and kaboom. Well, the feature would not be on by default. So careful evaluation and planning should prevent abuses. We generally use hosts now with plenty of RAM and the per-filesystem overhead for ZFS doesn't cause much concern. However, on a scratch box, try creating a big stack of filesystems - you can end up with a pool that consumes so much memory you can't import it! 'rmdir' question requires some thinking, my first reaction is it should do zfs destroy... .. which will fail if there's a snapshot, for example. The problem seems to be reasonably complex - compounded by the fact that many programs that create or remove directories do so directly - not by calling externals that would be ZFS aware. Well, snapshots could be destroyed too, nothing stops from doing that. BTW, I am not proposing to implement this feature in mkdir/rmdir userland utility, I am proposing to implement the feature in ZFS kernel code responsible for directory creation/removal. -- Andriy Gapon ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss