Re: [zfs-discuss] missing files on copy
On Jan 30, 2008 1:34 AM, Carson Gaspar [EMAIL PROTECTED] wrote: If this is Sun's cp, file a bug. It's failing to notice that it didn't provide a large enough buffer to getdents(), so it only got partial results. Of course, the getdents() API is rather unfortunate. It appears the only safe algorithm is: while ((r = getdents(...)) 0) { /* process results */ } if (r 0) { /* handle error */ } You _always_ have to call it at least twice to be sure you've gotten everything. In OpenSolaris, cp uses (indirectly) readdir(), not raw getdents(). http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/lib/libcmd/common/cp.c#487 which uses the build-a-linked-list code here: http://src.opensolaris.org/source/xref/sfw/usr/src/cmd/coreutils/coreutils-6.7/lib/fts.c#913 That code appears to error out and return incomplete results if a) the filename is too long or b) an integer overflows. Christopher's filenames are only 96 chars; could Unicode be involved somehow? b) seems unlikely in the extreme. It still seems like a bug, but I don't see where it is. I am only an egg ;-) Will ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
Christopher Gorski wrote: I noticed that the first calls in the cp and ls to getdents() return similar file lists, with the same values. However, in the ls, it makes a second call to getdents(): If this is Sun's cp, file a bug. It's failing to notice that it didn't provide a large enough buffer to getdents(), so it only got partial results. cp doesn't use getdents() but it uses readdir() instead; the whole buffer is hidden to it. Of course, the getdents() API is rather unfortunate. It appears the only safe algorithm is: while ((r = getdents(...)) 0) { /* process results */ } if (r 0) { /* handle error */ } You _always_ have to call it at least twice to be sure you've gotten everything. That's why you never use getdents but rather readdir() which hides this for you. It appears that the off_t of the directory entries in the particular second read is 2^32; so perhaps a cp which hasn't been compiled with handle large files is being used? Casper ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
That code appears to error out and return incomplete results if a) the filename is too long or b) an integer overflows. Christopher's filenames are only 96 chars; could Unicode be involved somehow? b) seems unlikely in the extreme. It still seems like a bug, but I don't see where it is. I am only an egg ;-) And ls would fail in the same manner. There's one piece of code in cp (see usr/src/cmd/mv/mv.c) which short-circuits a readdir-loop: while ((dp = readdir(srcdirp)) != NULL) { int ret; if ((ret = traverse_attrfile(dp, source, target, 1)) == -1) continue; else if (ret 0) { ++error; goto out; } This is strange to me because all other failures result in cp going over to the next file. Casper ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZIL controls in Solaris 10 U4?
Jonathan Loran writes: Is it true that Solaris 10 u4 does not have any of the nice ZIL controls that exist in the various recent Open Solaris flavors? I would like to move my ZIL to solid state storage, but I fear I can't do it until I have another update. Heck, I would be happy to just be able to turn the ZIL off to see how my NFS on ZFS performance is effected before spending the $'s. Anyone know when will we see this in Solaris 10? You can certainly turn it off with any release (Jim's link). It's true that S10u4 does not have the Separate Intent Log to allow using an SSD for ZIL blocks. I believe S10U5 will have that feature. As noted, disabling the ZIL won't lead to ZFS pool corruption, just DBcorruption(that includes NFS clients). To protect against that, in the event of a server crash with zil_disable=1, you'd need to reboot all NFS clients of the server (clear the client's caches) and better do this before the server comes back up (kind of a raw proposition here). -r Thanks, Jon -- - _/ _/ / - Jonathan Loran - - -/ / /IT Manager - - _ / _ / / Space Sciences Laboratory, UC Berkeley -/ / / (510) 643-5146 [EMAIL PROTECTED] - __/__/__/ AST:7731^29u18e3 ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
Will Murnane [EMAIL PROTECTED] wrote: On Jan 30, 2008 1:34 AM, Carson Gaspar [EMAIL PROTECTED] wrote: If this is Sun's cp, file a bug. It's failing to notice that it didn't provide a large enough buffer to getdents(), so it only got partial results. Of course, the getdents() API is rather unfortunate. It appears the only safe algorithm is: while ((r = getdents(...)) 0) { /* process results */ } if (r 0) { /* handle error */ } You _always_ have to call it at least twice to be sure you've gotten everything. In OpenSolaris, cp uses (indirectly) readdir(), not raw getdents(). http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/lib/libcmd/common/cp.c#487 which uses the build-a-linked-list code here: http://src.opensolaris.org/source/xref/sfw/usr/src/cmd/coreutils/coreutils-6.7/lib/fts.c#913 That code appears to error out and return incomplete results if a) the filename is too long or b) an integer overflows. Christopher's filenames are only 96 chars; could Unicode be involved somehow? b) seems unlikely in the extreme. It still seems like a bug, but I don't see where it is. I am only an egg ;-) An interesting thought We of course need to know whether the user used /bin/cp or a shadow implementation from ksh93. I did never see any problems with star(1) and star(1)/libfind(3) are heavy readdir(3) users... Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
Christopher Gorski [EMAIL PROTECTED] wrote: Of course, the getdents() API is rather unfortunate. It appears the only safe algorithm is: while ((r = getdents(...)) 0) { /* process results */ } if (r 0) { /* handle error */ } You _always_ have to call it at least twice to be sure you've gotten everything. Yes, it is Sun's cp. I'm trying, with some difficulty, to figure out exactly how to reproduce this error in a way not specific to my data. I copied a set of randomly generated files with a deep directory structure and cp seems to correctly call getdents() multiple times. Note that cp (mv) does not call getdents() directly but readdir(). If there is a problem, it is most likely in readdir() and it really looks strangee that ls(1) (although it uses the same implementaion) works for you. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
[EMAIL PROTECTED] wrote: And ls would fail in the same manner. There's one piece of code in cp (see usr/src/cmd/mv/mv.c) which short-circuits a readdir-loop: while ((dp = readdir(srcdirp)) != NULL) { int ret; if ((ret = traverse_attrfile(dp, source, target, 1)) == -1) continue; else if (ret 0) { ++error; goto out; } This is strange to me because all other failures result in cp going over to the next file. traverse_attrfile() returns -1 only for: if ((dp-d_name[0] == '.' dp-d_name[1] == '\0') || (dp-d_name[0] == '.' dp-d_name[1] == '.' dp-d_name[2] == '\0') || (sysattr_type(dp-d_name) == _RO_SATTR) || (sysattr_type(dp-d_name) == _RW_SATTR)) return (-1); So this primarily skips '.' and '..'. The rest seems to check for DOS extensions in extended attributes. but this is only done to copy attributes and not files. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
Hello Christopher, Wednesday, January 30, 2008, 7:27:01 AM, you wrote: CG Carson Gaspar wrote: Christopher Gorski wrote: I noticed that the first calls in the cp and ls to getdents() return similar file lists, with the same values. However, in the ls, it makes a second call to getdents(): If this is Sun's cp, file a bug. It's failing to notice that it didn't provide a large enough buffer to getdents(), so it only got partial results. Of course, the getdents() API is rather unfortunate. It appears the only safe algorithm is: while ((r = getdents(...)) 0) { /* process results */ } if (r 0) { /* handle error */ } You _always_ have to call it at least twice to be sure you've gotten everything. CG Yes, it is Sun's cp. I'm trying, with some difficulty, to figure out CG exactly how to reproduce this error in a way not specific to my data. I CG copied a set of randomly generated files with a deep directory structure CG and cp seems to correctly call getdents() multiple times. If you could re-create empty files - exactly the same directory atructure and file names, check if you still got a problem. If you do, then if you could send a script here (mkdir's -p and touch) so we can investigate. Assuming your file names and directory structure can be made public. -- Best regards, Robert Milkowskimailto:[EMAIL PROTECTED] http://milek.blogspot.com ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
Joerg Schilling wrote: Will Murnane [EMAIL PROTECTED] wrote: On Jan 30, 2008 1:34 AM, Carson Gaspar [EMAIL PROTECTED] wrote: If this is Sun's cp, file a bug. It's failing to notice that it didn't provide a large enough buffer to getdents(), so it only got partial results. Of course, the getdents() API is rather unfortunate. It appears the only safe algorithm is: while ((r = getdents(...)) 0) { /* process results */ } if (r 0) { /* handle error */ } You _always_ have to call it at least twice to be sure you've gotten everything. In OpenSolaris, cp uses (indirectly) readdir(), not raw getdents(). http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/lib/libcmd/common/cp.c#487 which uses the build-a-linked-list code here: http://src.opensolaris.org/source/xref/sfw/usr/src/cmd/coreutils/coreutils-6.7/lib/fts.c#913 That code appears to error out and return incomplete results if a) the filename is too long or b) an integer overflows. Christopher's filenames are only 96 chars; could Unicode be involved somehow? b) seems unlikely in the extreme. It still seems like a bug, but I don't see where it is. I am only an egg ;-) An interesting thought We of course need to know whether the user used /bin/cp or a shadow implementation from ksh93. I did never see any problems with star(1) and star(1)/libfind(3) are heavy readdir(3) users... Jörg I am able to replicate the problem in bash using: #truss -tall -vall -o /tmp/getdents.bin.cp.truss /bin/cp -pr /pond/photos/* /pond/copytestsame/ So I'm assuming that's using /bin/cp Also, from my _very limited_ investigation this morning, it seems that #grep Err /tmp/getdents.bin.cp.truss | grep -v ENOENT | grep getdents returns entries such as: getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEE34000, 8192) Err#9 EBADF ...(truncated) whereas it seems like with a copy where everything is transferred correctly that the above statement returns no getdents64() with an EBADF error, leading me to believe that somewhere along the line getdents64 is attempted to be called but that the descriptor is invalidated somehow. Again...I am only gleaming that from a very limited test. -Chris ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
Christopher Gorski [EMAIL PROTECTED] wrote: I am able to replicate the problem in bash using: #truss -tall -vall -o /tmp/getdents.bin.cp.truss /bin/cp -pr /pond/photos/* /pond/copytestsame/ So I'm assuming that's using /bin/cp Also, from my _very limited_ investigation this morning, it seems that #grep Err /tmp/getdents.bin.cp.truss | grep -v ENOENT | grep getdents returns entries such as: getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEE34000, 8192) Err#9 EBADF ...(truncated) If you get this, you may need to provide the full truss output to allow to understand what'ts happening. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
Robert Milkowski [EMAIL PROTECTED] wrote: If you could re-create empty files - exactly the same directory atructure and file names, check if you still got a problem. If you do, then if you could send a script here (mkdir's -p and touch) so we can investigate. If you like to replicate a long directory structure with empty files, you can use star: star -c -meta f=/tmp/x.tar -C dir . and later: star -xp -xmeta f=/tmp/x.tar Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] ZFS configuration for a thumper
Hi all I've a Sun X4500 with 48 disk of 750Go The server come with Solaris install on two disk. That's mean I've got 46 disk for ZFS. When I look the defautl configuration of the zpool zpool create -f zpool1 raidz c0t0d0 c1t0d0 c4t0d0 c6t0d0 c7t0d0 zpool add -f zpool1 raidz c0t1d0 c1t1d0 c4t1d0 c5t1d0 c6t1d0 c7t1d0 zpool add -f zpool1 raidz c0t2d0 c1t2d0 c4t2d0 c5t2d0 c6t2d0 c7t2d0 zpool add -f zpool1 raidz c0t3d0 c1t3d0 c4t3d0 c5t3d0 c6t3d0 c7t3d0 zpool add -f zpool1 raidz c0t4d0 c1t4d0 c4t4d0 c6t4d0 c7t4d0 zpool add -f zpool1 raidz c0t5d0 c1t5d0 c4t5d0 c5t5d0 c6t5d0 c7t5d0 zpool add -f zpool1 raidz c0t6d0 c1t6d0 c4t6d0 c5t6d0 c6t6d0 c7t6d0 zpool add -f zpool1 raidz c0t7d0 c1t7d0 c4t7d0 c5t7d0 c6t7d0 c7t7d0 that's mean there'are pool with 5 disk and other with 6 disk. When I want to do the same I've got this message : mismatched replication level: pool uses 5-way raidz and new vdev uses 6-way raidz I can force this with «-f» option. But what's that mean (sorry if the question is stupid). What's kind of pool you use with 46 disk ? (46=2*23 and 23 is prime number that's mean I can make raidz with 6 or 7 or any number of disk). Regards. -- Albert SHIH Observatoire de Paris Meudon SIO batiment 15 Heure local/Local time: Mer 30 jan 2008 16:36:49 CET ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
Also, from my _very limited_ investigation this morning, it seems tha= t #grep Err /tmp/getdents.bin.cp.truss | grep -v ENOENT | grep getdents returns entries such as: getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEE34000, 8192) Err#9 EBADF =2E..(truncated) Ah, this looks like someone closed stdin and then something weird happened. Hm. We need full truss out, specifically of all calls which return or release filedescriptors. The plot thickens. Casper ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS configuration for a thumper
Le 30/01/2008 à 11:01:35-0500, Kyle McDonald a écrit Albert Shih wrote: What's kind of pool you use with 46 disk ? (46=2*23 and 23 is prime number that's mean I can make raidz with 6 or 7 or any number of disk). Depending on needs for space vs. performance, I'd probably pixk eithr 5*9 or 9*5, with 1 hot spare. Thanks for the tips... How you can check the speed (I'm totally newbie on Solaris) I've use mkfile 10g for write and I've got same perf with 5*9 or 9*5. Have you some advice about tool like iozone ? Regards. -- Albert SHIH Observatoire de Paris Meudon SIO batiment 15 Heure local/Local time: Mer 30 jan 2008 17:10:55 CET ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS configuration for a thumper
On 1/30/08, Albert Shih [EMAIL PROTECTED] wrote: Thanks for the tips... How you can check the speed (I'm totally newbie on Solaris) I've use mkfile 10g for write and I've got same perf with 5*9 or 9*5. Have you some advice about tool like iozone ? Regards. -- Albert SHIH Observatoire de Paris Meudon SIO batiment 15 Heure local/Local time: Mer 30 jan 2008 17:10:55 CET ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss I'd take a look at bonnie++ http://www.sunfreeware.com/programlistintel10.html#bonnie++ --Tim ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS configuration for a thumper
Albert Shih wrote: What's kind of pool you use with 46 disk ? (46=2*23 and 23 is prime number that's mean I can make raidz with 6 or 7 or any number of disk). Depending on needs for space vs. performance, I'd probably pixk eithr 5*9 or 9*5, with 1 hot spare. -Kyle Regards. -- Albert SHIH Observatoire de Paris Meudon SIO batiment 15 Heure local/Local time: Mer 30 jan 2008 16:36:49 CET ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
Hello Joerg, Wednesday, January 30, 2008, 2:56:27 PM, you wrote: JS Robert Milkowski [EMAIL PROTECTED] wrote: If you could re-create empty files - exactly the same directory atructure and file names, check if you still got a problem. If you do, then if you could send a script here (mkdir's -p and touch) so we can investigate. JS If you like to replicate a long directory structure with empty files, JS you can use star: JS star -c -meta f=/tmp/x.tar -C dir . JS and later: JS star -xp -xmeta f=/tmp/x.tar It really is a swiss knife :) That's a handy one (although it's a first time I actually have seen a need for such functionality). -- Best regards, Robertmailto:[EMAIL PROTECTED] http://milek.blogspot.com ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] missing files on copy
[EMAIL PROTECTED] wrote: Also, from my _very limited_ investigation this morning, it seems tha= t #grep Err /tmp/getdents.bin.cp.truss | grep -v ENOENT | grep getdents returns entries such as: getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEC92000, 8192) Err#9 EBADF getdents64(0, 0xFEE34000, 8192) Err#9 EBADF =2E..(truncated) Ah, this looks like someone closed stdin and then something weird happened. Hm. stdin is usually not a directory ;-9 This looks much more weird Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZIL controls in Solaris 10 U4?
Roch - PAE wrote: Jonathan Loran writes: Is it true that Solaris 10 u4 does not have any of the nice ZIL controls that exist in the various recent Open Solaris flavors? I would like to move my ZIL to solid state storage, but I fear I can't do it until I have another update. Heck, I would be happy to just be able to turn the ZIL off to see how my NFS on ZFS performance is effected before spending the $'s. Anyone know when will we see this in Solaris 10? You can certainly turn it off with any release (Jim's link). It's true that S10u4 does not have the Separate Intent Log to allow using an SSD for ZIL blocks. I believe S10U5 will have that feature. Unfortunately it will not. A lot of ZFS fixes and features that had existed for a while will not be in U5 (for reasons I can't go into here). They should be in S10U6... Neil. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZIL controls in Solaris 10 U4?
Neil Perrin wrote: Roch - PAE wrote: Jonathan Loran writes: Is it true that Solaris 10 u4 does not have any of the nice ZIL controls that exist in the various recent Open Solaris flavors? I would like to move my ZIL to solid state storage, but I fear I can't do it until I have another update. Heck, I would be happy to just be able to turn the ZIL off to see how my NFS on ZFS performance is effected before spending the $'s. Anyone know when will we see this in Solaris 10? You can certainly turn it off with any release (Jim's link). It's true that S10u4 does not have the Separate Intent Log to allow using an SSD for ZIL blocks. I believe S10U5 will have that feature. Don't think we can live with this. Thanks Unfortunately it will not. A lot of ZFS fixes and features that had existed for a while will not be in U5 (for reasons I can't go into here). They should be in S10U6... Neil. I feel like we're being hung out to dry here. I've got 70TB on 9 various Solaris 10 u4 servers, with different data sets. All of these are NFS servers. Two servers have a ton of small files, with a lot of read and write updating, and NFS performance on these are abysmal. ZFS is installed on SAN array's (my first mistake). I will test by disabling the ZIL, but if it turns out the ZIL needs to be on a separate device, we're hosed. Before ranting any more, I'll do the test of disabling the ZIL. We may have to build out these systems with Open Solaris, but that will be hard as they are in production. I would have to install the new OS on test systems and swap out the drives during scheduled down time. Ouch. Jon -- - _/ _/ / - Jonathan Loran - - -/ / /IT Manager - - _ / _ / / Space Sciences Laboratory, UC Berkeley -/ / / (510) 643-5146 [EMAIL PROTECTED] - __/__/__/ AST:7731^29u18e3 ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZIL controls in Solaris 10 U4?
Are you already running with zfs_nocacheflush=1? We have SAN arrays with dual battery-backed controllers for the cache, so we definitely have this set on all our production systems. It makes a big difference for us. As I said before I don't see the catastrophe in disabling ZIL though. We actually run our production Cyrus mail servers using failover servers so our downtime is typically just the small interval to switch active idle nodes anyhow. We did this mainly for patching purposes. But we toyed with the idea of running OpenSolaris on them, then just upgrading the idle node to new OpenSolaris image every month using Jumpstart and switching to it. Anything goes wrong switch back to the other node. What we ended up doing, for political reasons, was putting the squeeze on our Sun reps and getting a 10u4 kernel spin patch with... what did they call it? Oh yeah a big wad of ZFS fixes. So this ends up being a hug PITA because for the next 6 months to a year we are tied to getting any kernel patches through this other channel rather than the usual way. But it does work for us, so there you are. Give my choice I'd go with OpenSolaris but that's a hard sell for datacenter management types. I think it's no big deal in a production shop with good JumpStart and CFengine setups, where any host should be rebuildable from scratch in a matter of hours. Good luck. This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS configuration for a thumper
[EMAIL PROTECTED] said: I'd take a look at bonnie++ http://www.sunfreeware.com/programlistintel10.html#bonnie++ Also filebench: http://www.solarisinternals.com/wiki/index.php/FileBench You'll see the most difference between 5x9 and 9x5 in small random reads: http://blogs.sun.com/relling/entry/zfs_raid_recommendations_space_performance http://blogs.sun.com/relling/entry/raid_recommendations_space_vs_mttdl http://lindsay.at/blog/archive/2007/04/15/zfs-performance-models-for-a-streamin g-server.html Regards, Marion ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZIL controls in Solaris 10 U4?
On Jan 30, 2008, at 3:44 PM, Vincent Fox wrote: What we ended up doing, for political reasons, was putting the squeeze on our Sun reps and getting a 10u4 kernel spin patch with... what did they call it? Oh yeah a big wad of ZFS fixes. So this ends up being a hug PITA because for the next 6 months to a year we are tied to getting any kernel patches through this other channel rather than the usual way. But it does work for us, so there you are. Speaking of big wad of ZFS fixes, is it me or is anyone else here getting kind of displeased over the glacial speed of the backporting of ZFS stability fixes to s10? It seems that we have to wait around 4-5 months for a oft-delayed s10 update for any fixes of substance to come out. Not only that, but also one day the zfs is its own patch, and then it is part of the current KU, and now it's part of the nfs patch where zfs isn't mentioned anywhere in the patch's synopsis. /dale ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZIL controls in Solaris 10 U4?
[EMAIL PROTECTED] said: I feel like we're being hung out to dry here. I've got 70TB on 9 various Solaris 10 u4 servers, with different data sets. All of these are NFS servers. Two servers have a ton of small files, with a lot of read and write updating, and NFS performance on these are abysmal. ZFS is installed on SAN array's (my first mistake). I will test by disabling the ZIL, but if it turns out the ZIL needs to be on a separate device, we're hosed. If you're using SAN arrays, you should be in good shape. I'll echo what Vincent Fox said about using either zfs_nocacheflush=1 (which is in S10U4), or setting the arrays to ignore the cache flush (SYNC_CACHE) requests. We do the latter here, and it makes a huge difference for NFS clients, basically putting the ZIL in NVRAM. However, I'm also unhappy about having to wait for S10U6 for the separate ZIL and/or cache features of ZFS. The lack of NV ZIL on our new Thumper makes it painfully slow over NFS for the large number of file create/delete type of workload. Here's a question: Would having the client mount with -o nocto have the same effect (for that particular client) as disabling the ZIL on the server? If so, it might be less drastic than losing the ZIL for everyone. Regards, Marion ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] ZFS under VMware
Hello, I'm planning to use VMware Server on Ubuntu to host multiple VMs, one of which will be a Solaris instance for the purposes of ZFS I would give the ZFS VM two physical disks for my zpool, e.g. /dev/sda and /dev/sdb, in addition to the VMware virtual disk for the Solaris OS Now I know that Solaris/ZFS likes to have total control over the disks to ensure writes are flushed as and when it is ready for them to happen, so I wonder if anybody comment on what implications using the disks in this way (i.e. through Linux and then VMware) has on the control Solaris has over these disks? By using a VM will I be missing out in terms of reliability? If so, can anybody suggest any improvements I could make while still allowing Solaris/ZFS to run in a VM? Many thanks, Lewis ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS under VMware
Lewis Thompson wrote: Hello, I'm planning to use VMware Server on Ubuntu to host multiple VMs, one of which will be a Solaris instance for the purposes of ZFS I would give the ZFS VM two physical disks for my zpool, e.g. /dev/sda and /dev/sdb, in addition to the VMware virtual disk for the Solaris OS Now I know that Solaris/ZFS likes to have total control over the disks to ensure writes are flushed as and when it is ready for them to happen, so I wonder if anybody comment on what implications using the disks in this way (i.e. through Linux and then VMware) has on the control Solaris has over these disks? By using a VM will I be missing out in terms of reliability? If so, can anybody suggest any improvements I could make while still allowing Solaris/ZFS to run in a VM? I'm not sure what the perf aspects would be but it depends on what the VMware software passes through. Does it ignore cache sync commands in its i/o stack? Got me. You won't be missing out on reliability but you will be introducing more layers in the stack where something could go wrong. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZIL controls in Solaris 10 U4?
Vincent Fox wrote: Are you already running with zfs_nocacheflush=1? We have SAN arrays with dual battery-backed controllers for the cache, so we definitely have this set on all our production systems. It makes a big difference for us. No, we're not using the zfs_nocacheflush=1, but our SAN array's are set to cache all writebacks, so it shouldn't be needed. I may test this, if I get the chance to reboot one of the servers, but I'll bet the storage arrays' are working correctly. As I said before I don't see the catastrophe in disabling ZIL though. No catastrophe, just a potential mess. We actually run our production Cyrus mail servers using failover servers so our downtime is typically just the small interval to switch active idle nodes anyhow. We did this mainly for patching purposes. Wish we could afford such replication. Poor EDU environment here, I'm afraid. But we toyed with the idea of running OpenSolaris on them, then just upgrading the idle node to new OpenSolaris image every month using Jumpstart and switching to it. Anything goes wrong switch back to the other node. What we ended up doing, for political reasons, was putting the squeeze on our Sun reps and getting a 10u4 kernel spin patch with... what did they call it? Oh yeah a big wad of ZFS fixes. So this ends up being a hug PITA because for the next 6 months to a year we are tied to getting any kernel patches through this other channel rather than the usual way. But it does work for us, so there you are. Mmmm, for us, Open Solaris may be easier. I manly was after stability, to be honest. Our ongoing experience with bleeding edge Linux is painful at times, and on our big iron, I want them to just work. but if they're so slow, they're not really working right, are they? Sigh... Give my choice I'd go with OpenSolaris but that's a hard sell for datacenter management types. I think it's no big deal in a production shop with good JumpStart and CFengine setups, where any host should be rebuildable from scratch in a matter of hours. Good luck. True, I'll think about that going forward. Thanks, Jon This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss -- - _/ _/ / - Jonathan Loran - - -/ / /IT Manager - - _ / _ / / Space Sciences Laboratory, UC Berkeley -/ / / (510) 643-5146 [EMAIL PROTECTED] - __/__/__/ AST:7731^29u18e3 ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZIL controls in Solaris 10 U4?
Jonathan Loran wrote: Vincent Fox wrote: Are you already running with zfs_nocacheflush=1? We have SAN arrays with dual battery-backed controllers for the cache, so we definitely have this set on all our production systems. It makes a big difference for us. No, we're not using the zfs_nocacheflush=1, but our SAN array's are set to cache all writebacks, so it shouldn't be needed. I may test this, if I get the chance to reboot one of the servers, but I'll bet the storage arrays' are working correctly. I think there's some confusion. ZFS and the ZIL issue controller commands to force the disk cache to be flushed to ensure data is on stable storage. If the disk cache is battery backed then the costly flush is unnecessary. As Vincent said, setting zfs_nocacheflush=1 can make a huge difference. Note that this is a system wide variable so all controllers serving ZFS devices should be non volatile to enable it. Neil. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZIL controls in Solaris 10 U4?
No, we're not using the zfs_nocacheflush=1, but our SAN array's are set to cache all writebacks, so it shouldn't be needed. I may test this, if get the chance to reboot one of the servers, but I'll bet the storage rrays' are working correctly. Bzzzt, wrong. Read up on a few threads about this variable. The ZFS flush command used equates to flush to rust for most any array. What this works out to, is your array is not using it's NV for what it's supposed to. You get a little data in the NV but it's tagged with this command that requires the NV to finish it's job and report back data is on disk, before proceeding. Hopefully at some point the array people and the ZFS people will have a meeting of the minds on this issue of having the array report to the OS yes I have battery-back SAFE NV and it will all just automagically work. Until then, we set the variable in /etc/system. This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] 30 seond hang, ls command....
I'm running Nevada build 81 on x86 on an Ultra 40. # uname -a SunOS zbit 5.11 snv_81 i86pc i386 i86pc Memory size: 8191 Megabytes I started with this zfs pool many dozens of builds ago, approx a year ago. I do live upgrade and zfs upgrade every few builds. When I have not accessed the zfs file systems for a long time, if I cd there and do an ls command, nothing happens for approx 30 seconds. Any clues how I would find out what is wrong? -- # zpool status -v pool: tank state: ONLINE scrub: none requested config: NAMESTATE READ WRITE CKSUM tankONLINE 0 0 0 raidz2ONLINE 0 0 0 c2d0ONLINE 0 0 0 c3d0ONLINE 0 0 0 c4d0ONLINE 0 0 0 c5d0ONLINE 0 0 0 c6d0ONLINE 0 0 0 c7d0ONLINE 0 0 0 c8d0ONLINE 0 0 0 errors: No known data errors # zfs list NAME USED AVAIL REFER MOUNTPOINT tank 172G 2.04T 52.3K /tank tank/arc 172G 2.04T 172G /zfs/arc # zpool list NAME SIZE USED AVAILCAP HEALTH ALTROOT tank 3.16T 242G 2.92T 7% ONLINE - ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] 30 seond hang, ls command....
Any chance the disks are being powered down, and you are waiting for them to power back up? Nathan. :) Neal Pollack wrote: I'm running Nevada build 81 on x86 on an Ultra 40. # uname -a SunOS zbit 5.11 snv_81 i86pc i386 i86pc Memory size: 8191 Megabytes I started with this zfs pool many dozens of builds ago, approx a year ago. I do live upgrade and zfs upgrade every few builds. When I have not accessed the zfs file systems for a long time, if I cd there and do an ls command, nothing happens for approx 30 seconds. Any clues how I would find out what is wrong? -- # zpool status -v pool: tank state: ONLINE scrub: none requested config: NAMESTATE READ WRITE CKSUM tankONLINE 0 0 0 raidz2ONLINE 0 0 0 c2d0ONLINE 0 0 0 c3d0ONLINE 0 0 0 c4d0ONLINE 0 0 0 c5d0ONLINE 0 0 0 c6d0ONLINE 0 0 0 c7d0ONLINE 0 0 0 c8d0ONLINE 0 0 0 errors: No known data errors # zfs list NAME USED AVAIL REFER MOUNTPOINT tank 172G 2.04T 52.3K /tank tank/arc 172G 2.04T 172G /zfs/arc # zpool list NAME SIZE USED AVAILCAP HEALTH ALTROOT tank 3.16T 242G 2.92T 7% ONLINE - ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] I.O error: zpool metadata corrupted after powercut
Last 2 weeks we had 2 zpools corrupted. Pool was visible via zpool import, but could not be imported anymore. During import attempt we got I/O error, After a first powercut we lost our jumpstart/nfsroot zpool (another pool was still OK). Luckaly jumpstart data was backed up and easely restored, nfsroot Filesystems where not but those where just test machines. We thought the metadata corruption was caused because of the zfs no cache flush setting we had configured in /etc/system (for perfomance reason) in combination with a non battery backuppped NVRAM cache (areca raid controller). zpool was raidz with 10 local sata disks (JBOD mode) 2 days ago we had another powercut in our test labo :-( And again one pool was lost. This system was not configured with zfs no cache flush. On the pool we had +/- 40 zvols used by running vm's (iscsi boot/swap/data disks for xen virtual box guests) The first failure was on a b68 system, the second on a b77 system. Last zpool was using iscsi disks: setup: pool mirror: iscsidisk1 san1 iscsidisk1 san2 mirror: iscsidisk2 san1 iscsidisk2 san2 I thought zfs was always persistent on disk, but apparently a power cut has can cause unrecoverable damage. I can accept the first failure (because of the dangerous setting), but loosing that second pool was unacceptable for me. Since no fsck alike utility is available for zfs I was wondering if there are any plans to create something like meta data repair tools? Using ZFS now for almost 1 year I was a big Fan, In one year I lost not 1 zpool till last week. At this time I'm concidering to say ZFS is not yet production ready any comment welcome... krdoor This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] Can't offline second disk in a mirror
Since I spend a lot of time going from machine to machine so I thought I'd carry a pool with me on a couple of USB keys. It all works fine but it's slow, so I thought I'd attach a file vdev to the pool and then offline the USB devices for speed, then undo when I want to take the keys with me. Unfortunately, it seems that once I've offlined one device, the mirror is marked as degraded and then I'm not allows to take the other USB key offline: # zpool create usb mirror /dev/dsk/c5t0d0p0 /dev/dsk/c6t0d0p0 # mkfile 2g /file # zpool attach usb c6t0d0p0 /file # zpool status pool: usb state: ONLINE scrub: resilver completed with 0 errors on Thu Jan 31 13:24:22 2008 config: NAME STATE READ WRITE CKSUM usb ONLINE 0 0 0 mirror ONLINE 0 0 0 c5t0d0p0 ONLINE 0 0 0 c6t0d0p0 ONLINE 0 0 0 /file ONLINE 0 0 0 errors: No known data errors # zpool offline usb c5t0d0p0 Bringing device c5t0d0p0 offline # zpool status pool: usb state: DEGRADED status: One or more devices has been taken offline by the administrator. Sufficient replicas exist for the pool to continue functioning in a degraded state. action: Online the device using 'zpool online' or replace the device with 'zpool replace'. scrub: resilver completed with 0 errors on Thu Jan 31 13:24:22 2008 config: NAME STATE READ WRITE CKSUM usb DEGRADED 0 0 0 mirror DEGRADED 0 0 0 c5t0d0p0 OFFLINE 0 0 0 c6t0d0p0 ONLINE 0 0 0 /file ONLINE 0 0 0 errors: No known data errors # zpool offline usb c6t0d0p0 cannot offline c6t0d0p0: no valid replicas # cat /etc/release Solaris 10 8/07 s10x_u4wos_12b X86 Copyright 2007 Sun Microsystems, Inc. All Rights Reserved. Use is subject to license terms. Assembled 16 August 2007 I've experimented with other configurations (not just keys and files, but slices as well) and found the same thing - once one device in a mirror is offline I can't offline any others, even though there are other (sometimes multiple) copies left. Of course, I can detach the device, but I was hoping to avoid a full resilver when I reattach. Is this the expected behaviour? Am I missing something that would mean that what I'm trying to do is a bad idea? Boyd ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] I.O error: zpool metadata corrupted after powercut
kristof wrote: Last 2 weeks we had 2 zpools corrupted. Pool was visible via zpool import, but could not be imported anymore. During import attempt we got I/O error, What exactly was the error message? Also look at the fma messages, as they are often more precise. -- richard After a first powercut we lost our jumpstart/nfsroot zpool (another pool was still OK). Luckaly jumpstart data was backed up and easely restored, nfsroot Filesystems where not but those where just test machines. We thought the metadata corruption was caused because of the zfs no cache flush setting we had configured in /etc/system (for perfomance reason) in combination with a non battery backuppped NVRAM cache (areca raid controller). zpool was raidz with 10 local sata disks (JBOD mode) 2 days ago we had another powercut in our test labo :-( And again one pool was lost. This system was not configured with zfs no cache flush. On the pool we had +/- 40 zvols used by running vm's (iscsi boot/swap/data disks for xen virtual box guests) The first failure was on a b68 system, the second on a b77 system. Last zpool was using iscsi disks: setup: pool mirror: iscsidisk1 san1 iscsidisk1 san2 mirror: iscsidisk2 san1 iscsidisk2 san2 I thought zfs was always persistent on disk, but apparently a power cut has can cause unrecoverable damage. I can accept the first failure (because of the dangerous setting), but loosing that second pool was unacceptable for me. Since no fsck alike utility is available for zfs I was wondering if there are any plans to create something like meta data repair tools? Using ZFS now for almost 1 year I was a big Fan, In one year I lost not 1 zpool till last week. At this time I'm concidering to say ZFS is not yet production ready any comment welcome... krdoor This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] Hardware RAID vs. ZFS RAID
Hello, I have a Dell 2950 with a Perc 5/i, two 300GB 15K SAS drives in a RAID0 array. I am considering going to ZFS and I would like to get some feedback about which situation would yield the highest performance: using the Perc 5/i to provide a hardware RAID0 that is presented as a single volume to OpenSolaris, or using the drives separately and creating the RAID0 with OpenSolaris and ZFS? Or maybe just adding the hardware RAID0 to a ZFS pool? Can anyone suggest some articles or FAQs on implementing ZFS RAID? Which situation would provide the highest read and write throughput? Thanks in advance This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] seen on freebsd-stable: reproducible zfs panic
Hi everybody, Greg pointed me to http://lists.freebsd.org/pipermail/freebsd-stable/2008-January/040136.html from a Daniel Eriksson: If you import and export more than one zpool FreeBSD will panic during shutdown. This bug is present in both RELENG_7 and RELENG_7_0 (I have not tested CURRENT). kgdb output: Syncing disks, vnodes remaining...2 1 0 0 done All buffers synced. vput: negative ref count 0xc2ad1aa0: tag ufs, type VDIR usecount 0, writecount 0, refcount 2 mountedhere 0 flags (VV_ROOT) VI_LOCKedv_object 0xc1030174 ref 0 pages 1 lock type ufs: EXCL (count 1) by thread 0xc296 (pid 1) ino 2, on dev ad0s1a panic: vput: negative ref cnt KDB: stack backtrace: db_trace_self_wrapper(c086ad8a,d3b19b68,c06265ba,c0868fd5,c08e9ca0,...) at db_trace_self_wrapper+0x26 kdb_backtrace(c0868fd5,c08e9ca0,c086f57a,d3b19b74,d3b19b74,...) at kdb_backtrace+0x29 panic(c086f57a,c085,c086f561,c2ad1aa0,d3b19b90,...) at panic+0xaa vput(c2ad1aa0,2,d3b19bf0,c296,c086eedd,...) at vput+0xdb dounmount(c2ba6d0c,8,c296,0,0,...) at dounmount+0x49f vfs_unmountall(c0868ebb,0,c2967000,8,d3b19c50,...) at vfs_unmountall+0x33 boot(c296,8,1,c295e000,c296,...) at boot+0x3e3 reboot(c296,d3b19cfc,4,c086b882,56,...) at reboot+0x66 syscall(d3b19d38) at syscall+0x33a Xint0x80_syscall() at Xint0x80_syscall+0x20 --- syscall (55, FreeBSD ELF32, reboot), eip = 0x8050903, esp = 0xbfbfe90c, ebp = 0xbfbfe9d8 --- Uptime: 2m42s Physical memory: 503 MB Dumping 39 MB: 24 8 Run this script and then reboot the computer to trigger the panic: dd if=/dev/zero of=/usr/_disk1 bs=1m count=80 dd if=/dev/zero of=/usr/_disk2 bs=1m count=80 mdconfig -f /usr/_disk1 -u 1 mdconfig -f /usr/_disk2 -u 2 /etc/rc.d/zfs forcestart zpool create tank1 md1 zpool create tank2 md2 sleep 2 touch /tank1/testfile touch /tank2/testfile sleep 2 zpool export tank2 zpool export tank1 sleep 10 zpool import tank1 zpool import tank2 sleep 2 touch /tank1/testfile touch /tank2/testfile sleep 2 zpool export tank2 zpool export tank1 /etc/rc.d/zfs forcestop sleep 2 mdconfig -d -u 1 mdconfig -d -u 2 rm /usr/_disk1 rm /usr/_disk2 /Daniel Eriksson Anybody seen anything like this, on Solaris or freebsd? I've got my doubts about whether Daniel's got a valid test. thanks, James C. McPherson -- Senior Kernel Software Engineer, Solaris Sun Microsystems http://blogs.sun.com/jmcp http://www.jmcp.homeunix.com/blog ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZIL controls in Solaris 10 U4?
However, I'm also unhappy about having to wait for S10U6 for the separate ZIL and/or cache features of ZFS. The lack of NV ZIL on our new Thumper makes it painfully slow over NFS for the large number of file create/delete type of workload. I did a bit of testing on this (because I'm in the same boat) and was able to work around it by breaking my filesystem up into lots of individual zfs filesystems. Although the performance of each one isn't great, as long as your load is threaded and distributed across filesystems, it should balance out. Steve Hillman Simon Fraser University This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss