Re: [zfs-discuss] What to do with a disk partition
Blake blake.ir...@gmail.com writes: I think you will be helped by looking at this document: http://www.solarisinternals.com/wiki/index.php/ZFS_Troubleshooting_Guide#ZFS_Root_Pool_Recommendations_and_Requirements It addresses many of your questions. I think the easiest way to back up your OS might be to attach a disk to the rpool as a mirror, use 'installgrub' to get the grub boot blocks onto the new mirror disk, then detach this disk and put it in storage. Thanks for the pointer... looks like I can confuse myself with this guide for quite a while ... hehe. One question springs to mind immediately about attaching a mirror for backup of the os. I'm doing this on pc hardware. I don't know enough yet to understand how zpools repair themselves or how the parity data really works to recreate missing data. Or what might happen if a zpool were mounted with a disk missing. Imagine I have all hardware controller plugins used up with disks in various pools. So to install a disk to use for the mirror, something else will have to be unhooked (I mean manually). When I boot up to transfer zpool data to the newly added mirror disk, one or another zpool will be missing a disk. Is that something that will cause some kind of big problem? Or would I be able to do something so the effected pool did'nt get mounted... or maybe other choices I have no idea about yet? ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] ACL interpretation
On page 202 of the December 2008 Solaris ZFS Administration Guide, it says the ACLs are processed in order. Then it says that an explicit allow ends processing (or at least it says that a later deny can't override an earlier allow). But that's all it says; it doesn't really describe the interpretation process completely. I certainly couldn't implement it from this! And I can't figure out what my ACLs should mean from this. In particular, does a matching deny entry also halt processing? Or does processing continue, meaning that a later allow can override an earlier deny? -- David Dyer-Bennet, d...@dd-b.net; http://dd-b.net/ Snapshots: http://dd-b.net/dd-b/SnapshotAlbum/data/ Photos: http://dd-b.net/photography/gallery/ Dragaera: http://dragaera.info ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] After createing zpool of combined 750gb only 229 shows
Summary: I'm doing something wrong here but not sure what. Put a 250gb and 500gb disk into zpool but only 229gb is available I have a 250 gb disk and a 500gb disk installed and configured as raidz1. both have efi labels when viewed with format/fdisk. fdisk c3d1: Total disk size is 30401 cylinders Cylinder size is 16065 (512 byte) blocks Cylinders Partition StatusType Start End Length% = == = === == === 1 EFI 0 3040130402100 fdisk c4d0: Total disk size is 60800 cylinders Cylinder size is 16065 (512 byte) blocks Cylinders Partition StatusType Start End Length% = == = === == === 1 EFI 0 6080060801100 After creating the zpool with: zpool create zbk raidz1 c3d1(250gb) c4d0(500gb) (with no errors) zpool status zbk pool: zbk state: ONLINE scrub: none requested config: NAMESTATE READ WRITE CKSUM zbk ONLINE 0 0 0 raidz1ONLINE 0 0 0 c3d1ONLINE 0 0 0 c4d0ONLINE 0 0 0 errors: No known data errors But then df -h shows only 229gb available df -h /zbk FilesystemSize Used Avail Use% Mounted on zbk 229G 22K 229G 1% /zbk ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Can VirtualBox run a 64 bit guests on 32 bit host
On Sat, Feb 28, 2009 at 01:20:54AM -0600, Harry Putnam wrote: So cutting to the chase here... would you happen to have a recommendation from your own experience, or something you've heard will work and that can stand more ram... my current setup tops out at 3gb. The link to the HCL that was posted is probably your best bet. I know very little about PC hardware as I've always worked on SPARC (or POWER/MIPS/etc). -brian -- Coding in C is like sending a 3 year old to do groceries. You gotta tell them exactly what you want or you'll end up with a cupboard full of pop tarts and pancake mix. -- IRC User (http://www.bash.org/?841435) ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] Freezing OpenSolaris with ZFS
Hi there, we set up an OpenSolaris/ZFS based storage server with two zpools: rpool is a mirror for the operating system. tank is a raidz for data storage. The system is used to store large video files and has attached 12x1GB SATA-drives (2 mirrored for the system). Everytime large files are copied around the system hangs without apparent reason, 50% kernel CPU usage (so one core is occupied totally) and about 2GB of free RAM (8GB installed). On idle nothing crashes. Furthermore every scrub on tank hangs the system up below 1% finished. Neither the /var/adm/messages nor the /var/log/syslog file contains any errors or warnings. We limited the ZFS ARC cache to 4GB with an entry in /etc/system. Does anyone has an idea what's happening there and how to solve the problem? Below some outputs which may help. Thanks and greetings from germany, Markus Denhoff, Sebastian Friederichs # zpool status tank pool: tank state: ONLINE scrub: none requested config: NAME STATE READ WRITE CKSUM tank ONLINE 0 0 0 raidz1 ONLINE 0 0 0 c6t2d0 ONLINE 0 0 0 c6t3d0 ONLINE 0 0 0 c6t4d0 ONLINE 0 0 0 c6t5d0 ONLINE 0 0 0 c6t6d0 ONLINE 0 0 0 c6t7d0 ONLINE 0 0 0 c6t8d0 ONLINE 0 0 0 c6t9d0 ONLINE 0 0 0 c6t10d0 ONLINE 0 0 0 c6t11d0 ONLINE 0 0 0 errors: No known data errors # zpool iostat capacity operationsbandwidth pool used avail read write read write -- - - - - - - rpool 37.8G 890G 3 2 94.7K 17.4K tank2.03T 7.03T112 0 4.62M906 -- - - - - - - # zfs list NAME USED AVAIL REFER MOUNTPOINT rpool 39.8G 874G72K /rpool rpool/ROOT35.7G 874G18K legacy rpool/ROOT/opensolaris35.6G 874G 35.3G / rpool/ROOT/opensolaris-1 89.9M 874G 2.47G /tmp/tmp8CN5TR rpool/dump2.00G 874G 2.00G - rpool/export 172M 874G19K /export rpool/export/home 172M 874G21K /export/home rpool/swap2.00G 876G24K - tank 1.81T 6.17T 32.2K /tank tank/data 1.81T 6.17T 1.77T /data tank/public-share 34.9K 6.17T 34.9K /public-share ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] After createing zpool of combined 750gb only 229 shows
On 15 March, 2009 - Harry Putnam sent me these 1,7K bytes: Summary: I'm doing something wrong here but not sure what. Put a 250gb and 500gb disk into zpool but only 229gb is available I have a 250 gb disk and a 500gb disk installed and configured as raidz1. both have efi labels when viewed with format/fdisk. fdisk c3d1: Total disk size is 30401 cylinders Cylinder size is 16065 (512 byte) blocks Cylinders Partition StatusType Start End Length% = == = === == === 1 EFI 0 3040130402100 fdisk c4d0: Total disk size is 60800 cylinders Cylinder size is 16065 (512 byte) blocks Cylinders Partition StatusType Start End Length% = == = === == === 1 EFI 0 6080060801100 After creating the zpool with: zpool create zbk raidz1 c3d1(250gb) c4d0(500gb) (with no errors) zpool status zbk pool: zbk state: ONLINE scrub: none requested config: NAMESTATE READ WRITE CKSUM zbk ONLINE 0 0 0 raidz1ONLINE 0 0 0 c3d1ONLINE 0 0 0 c4d0ONLINE 0 0 0 errors: No known data errors But then df -h shows only 229gb available df -h /zbk FilesystemSize Used Avail Use% Mounted on zbk 229G 22K 229G 1% /zbk You are using raidz1 (which is less useful when you just have 2 disks, use a mirror instead for same safety but higher performance) which can't use more space than the smallest disk.. so half of the larger disk is unused in this case. 250 vs 229GB is 1000 vs 1024 when counting MB/kB etc. For redundant storage, use disks of the same size.. If you just want to add up both disks together (and lose all data if one disk goes belly up), create it with: zpool create zkb c3d1 c4d0 I seem to recall that ZFS should complain and point out that they are of (major) different size.. But apparently not.. /Tomas -- Tomas Ögren, st...@acc.umu.se, http://www.acc.umu.se/~stric/ |- Student at Computing Science, University of Umeå `- Sysadmin at {cs,acc}.umu.se ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] After createing zpool of combined 750gb only 229 shows
Tomas Ögren st...@acc.umu.se writes: I seem to recall that ZFS should complain and point out that they are of (major) different size.. But apparently not.. Thanks for the tips. It actually did complain about size difference at one point. I used the -f option. But later destroyed the zpool I'd created that way. I thought the problem had something to do with fdisk partitions so fdisked and deleted all partitions. (There was 1 on each drive). After that when I created the raidz1 with those two there was no complaint so I thought I was walking in tall cotton. Only later when I ran df -h and saw that zfs had reduced the zpool to lowest disk size did I realize I didn't really understand what I was doing. Your tips summarize what I'd already found on line before seeing your post. Very useful to hear from experienced user on this. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] CLI grinds to a halt during backups
Here is the info required. PID USERNAME SIZE RSS STATE PRI NICE TIME CPU PROCESS/NLWP 6179 nobody312M 225M sleep 510 12:42:09 0.8% BackupPC_dump/1 7783 root 3812K 2984K cpu7500 0:00:03 0.4% prstat/1 7803 root 2948K 1736K sleep 540 0:00:00 0.0% top/1 900 nobody 88M 4140K cpu3590 0:00:00 0.0% httpd/1 832 nobody 88M 3800K sleep 590 0:00:00 0.0% httpd/1 898 nobody 88M 3700K sleep 590 0:00:00 0.0% httpd/1 7782 root 6172K 3448K sleep 590 0:00:00 0.0% sshd/1 7772 root 2748K 1644K sleep 590 0:00:00 0.0% iostat/1 746 root 3164K 1616K sleep 590 0:00:00 0.0% dmispd/1 516 root 2800K 1532K sleep 590 0:00:00 0.0% automountd/2 513 root 2516K 948K sleep 590 0:00:00 0.0% automountd/2 532 root 4120K 1876K sleep 590 0:00:00 0.0% syslogd/13 829 nobody 88M 3568K sleep 590 0:00:00 0.0% httpd/1 831 nobody 88M 4124K sleep 590 0:00:00 0.0% httpd/1 352 daemon 2436K 1292K sleep 60 -20 0:00:00 0.0% nfs4cbd/2 430 root 2060K 676K sleep 590 0:00:00 0.0% smcboot/1 300 root 2752K 940K sleep 590 0:00:00 0.0% cron/1 359 daemon 4704K 1752K sleep 590 0:00:00 0.0% nfsmapid/3 173 daemon 4216K 2068K sleep 590 0:00:00 0.0% kcfd/3 517 root 3020K 2020K sleep 590 0:00:00 0.0% vold/5 152 root 1820K 1028K sleep 590 0:00:00 0.0% powerd/3 425 root 4884K 3260K sleep 590 0:00:00 0.0% inetd/3 138 root 4964K 1908K sleep 590 0:00:00 0.0% syseventd/15 428 root 2060K 964K sleep 590 0:00:00 0.0% smcboot/1 393 root 2068K 912K sleep 590 0:00:00 0.0% sac/1 163 root 3684K 2000K sleep 590 0:00:00 0.0% devfsadm/6 167 root 3880K 2620K sleep 590 0:00:00 0.0% picld/5 899 nobody 88M 4100K sleep 590 0:00:00 0.0% httpd/1 398 root 1428K 648K sleep 590 0:00:00 0.0% utmpd/1 350 daemon 2768K 1592K sleep 590 0:00:00 0.0% statd/1 NPROC USERNAME SWAP RSS MEMORY TIME CPU 12 nobody901M 512M 6.2% 12:46:35 0.8% 47 root 329M 209M 2.5% 0:14:01 0.4% 1 noaccess 171M 204M 2.5% 0:00:59 0.0% 1 smmsp1200K 3272K 0.0% 0:00:00 0.0% 6 daemon 6352K 6216K 0.1% 0:00:00 0.0% Total: 67 processes, 243 lwps, load averages: 18.49, 15.84, 13.77 iostat -x 5 extended device statisticsdevicer/s w/s kr/s kw/s wait actv svc_t %w %b sd0 0.00.00.00.0 0.0 0.00.0 0 0 sd1 0.00.00.00.0 0.0 0.00.0 0 0 sd2 0.0 18.90.0 195.9 0.0 0.01.2 0 1 sd3 0.0 19.40.0 196.4 0.0 0.01.4 0 1 sd4 0.00.00.00.0 0.0 0.00.0 0 0 sd5 0.0 18.90.0 176.4 0.0 0.01.3 0 1 sd6 0.0 18.40.0 166.2 0.0 0.01.4 0 1 sd7 0.0 19.40.0 175.7 0.0 0.01.3 0 1 sd8 0.0 20.20.0 178.3 0.0 0.01.3 0 1 sd9 0.0 19.90.0 213.8 0.0 0.01.1 0 1 sd10 0.0 19.40.0 196.5 0.0 0.01.2 0 1 sd11 0.0 19.70.0 200.6 0.0 0.01.2 0 1 sd12 0.0 19.40.0 175.9 0.0 0.01.4 0 1 sd13 0.0 19.40.0 188.0 0.0 0.01.3 0 1 nfs1 0.00.00.00.0 0.0 0.00.0 0 0 zpool iostat 5 (if you are using ZFS) -bash-3.00# zpool iostat 5 capacity operationsbandwidth pool used avail read write read write -- - - - - - - pool1 1.68T 8.32T 3168 371K 9.81M pool1 1.68T 8.32T 0 68 0 1.58M pool1 1.68T 8.32T 0 98 0 2.29M pool1 1.68T 8.32T 0 36 0 1.23M pool1 1.68T 8.32T 0103 0 2.67M pool1 1.68T 8.32T 0 16 0 90.8K pool1 1.68T 8.32T 0104 0 2.88M pool1 1.68T 8.32T 0 86 0 1.65M pool1 1.68T 8.32T 0 35 0 1.03M pool1 1.68T 8.32T 0162 0 4.03M pool1 1.68T 8.32T 0 46 0 1.35M pool1 1.68T 8.32T 0 53 0 1.11M pool1 1.68T 8.32T 0 75 0 2.15M Also top: last pid: 7803; load avg: 18.5, 15.8, 13.8; up 1+21:19:03 10:06:00 67 processes: 63 sleeping, 2 running, 2 on cpu CPU states: 7.1% idle, 0.6% user, 92.3% kernel, 0.0% iowait, 0.0% swap Kernel: 194 ctxsw, 13 trap, 18419 intr, 2955 syscall, 9 flt Memory: 8191M phys mem, 615M free mem, 20G total swap, 20G free swap PID USERNAME LWP PRI NICE SIZE RES STATETIMECPU COMMAND 7783 root 1 500 3812K 2984K run 0:03 0.70% prstat 6179 nobody 1 510 312M 225M run762:09 0.48% BackupPC_dump 898 nobody 1 590 88M 3700K
Re: [zfs-discuss] ACL interpretation
David Dyer-Bennet wrote: On page 202 of the December 2008 Solaris ZFS Administration Guide, it says the ACLs are processed in order. Then it says that an explicit allow ends processing (or at least it says that a later deny can't override an earlier allow). But that's all it says; it doesn't really describe the interpretation process completely. I certainly couldn't implement it from this! And I can't figure out what my ACLs should mean from this. In particular, does a matching deny entry also halt processing? Or does processing continue, meaning that a later allow can override an earlier deny? An ACL is processed from top to bottom. A deny entry can't take away an already granted allow nor can a allow take away an denied deny entry. For example: user:joe:read_data/write_data:allow user:joe:write_data:deny In this case joe would be allowed read_data and write_data whereas user:joe:write_data/execute:deny user:joe:read_data/write_data:allow would deny joe the ability to execute or write_data, but joe could still read the files data. Once a bit has been denied only a privilege subsystem override can give you that ability. -Mark ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ACL interpretation
Mark Shellenbaum wrote: David Dyer-Bennet wrote: On page 202 of the December 2008 Solaris ZFS Administration Guide, it says the ACLs are processed in order. Then it says that an explicit allow ends processing (or at least it says that a later deny can't override an earlier allow). But that's all it says; it doesn't really describe the interpretation process completely. I certainly couldn't implement it from this! And I can't figure out what my ACLs should mean from this. In particular, does a matching deny entry also halt processing? Or does processing continue, meaning that a later allow can override an earlier deny? An ACL is processed from top to bottom. A deny entry can't take away an already granted allow nor can a allow take away an denied deny entry. For example: [snip] Once a bit has been denied only a privilege subsystem override can give you that ability. Thanks, that's what I guessed and what simple experiments seemed to show, but Happy to have it confirmed. So the list is processed top to bottom and the first definite answer is THE answer. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Freezing OpenSolaris with ZFS
This sounds quite like the problems I've been having with a spotty sata controller and/or motherboard. See my thread from last week about copying large amounts of data that forced a reboot. Lots of good info from engineers and users in that thread. On Sun, Mar 15, 2009 at 1:17 PM, Markus Denhoff denh...@net-bite.net wrote: Hi there, we set up an OpenSolaris/ZFS based storage server with two zpools: rpool is a mirror for the operating system. tank is a raidz for data storage. The system is used to store large video files and has attached 12x1GB SATA-drives (2 mirrored for the system). Everytime large files are copied around the system hangs without apparent reason, 50% kernel CPU usage (so one core is occupied totally) and about 2GB of free RAM (8GB installed). On idle nothing crashes. Furthermore every scrub on tank hangs the system up below 1% finished. Neither the /var/adm/messages nor the /var/log/syslog file contains any errors or warnings. We limited the ZFS ARC cache to 4GB with an entry in /etc/system. Does anyone has an idea what's happening there and how to solve the problem? Below some outputs which may help. Thanks and greetings from germany, Markus Denhoff, Sebastian Friederichs # zpool status tank pool: tank state: ONLINE scrub: none requested config: NAME STATE READ WRITE CKSUM tank ONLINE 0 0 0 raidz1 ONLINE 0 0 0 c6t2d0 ONLINE 0 0 0 c6t3d0 ONLINE 0 0 0 c6t4d0 ONLINE 0 0 0 c6t5d0 ONLINE 0 0 0 c6t6d0 ONLINE 0 0 0 c6t7d0 ONLINE 0 0 0 c6t8d0 ONLINE 0 0 0 c6t9d0 ONLINE 0 0 0 c6t10d0 ONLINE 0 0 0 c6t11d0 ONLINE 0 0 0 errors: No known data errors # zpool iostat capacity operations bandwidth pool used avail read write read write -- - - - - - - rpool 37.8G 890G 3 2 94.7K 17.4K tank 2.03T 7.03T 112 0 4.62M 906 -- - - - - - - # zfs list NAME USED AVAIL REFER MOUNTPOINT rpool 39.8G 874G 72K /rpool rpool/ROOT 35.7G 874G 18K legacy rpool/ROOT/opensolaris 35.6G 874G 35.3G / rpool/ROOT/opensolaris-1 89.9M 874G 2.47G /tmp/tmp8CN5TR rpool/dump 2.00G 874G 2.00G - rpool/export 172M 874G 19K /export rpool/export/home 172M 874G 21K /export/home rpool/swap 2.00G 876G 24K - tank 1.81T 6.17T 32.2K /tank tank/data 1.81T 6.17T 1.77T /data tank/public-share 34.9K 6.17T 34.9K /public-share ___ 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] Freezing OpenSolaris with ZFS
On Sun, Mar 15, 2009 at 6:42 PM, Blake blake.ir...@gmail.com wrote: On Sun, Mar 15, 2009 at 1:17 PM, Markus Denhoff denh...@net-bite.net wrote: Hi there, we set up an OpenSolaris/ZFS based storage server with two zpools: rpool is a mirror for the operating system. tank is a raidz for data storage. The system is used to store large video files and has attached 12x1GB SATA-drives (2 mirrored for the system). Everytime large files are copied around the system hangs without apparent reason, 50% kernel CPU usage (so one core is occupied totally) and about 2GB of free RAM (8GB installed). On idle nothing crashes. Furthermore every scrub on tank hangs the system up below 1% finished. Neither the /var/adm/messages nor the /var/log/syslog file contains any errors or warnings. We limited the ZFS ARC cache to 4GB with an entry in /etc/system. Does anyone has an idea what's happening there and how to solve the problem? Below some outputs which may help. Thanks and greetings from germany, Markus Denhoff, Sebastian Friederichs # zpool status tank pool: tank state: ONLINE scrub: none requested config: NAME STATE READ WRITE CKSUM tank ONLINE 0 0 0 raidz1 ONLINE 0 0 0 c6t2d0 ONLINE 0 0 0 c6t3d0 ONLINE 0 0 0 c6t4d0 ONLINE 0 0 0 c6t5d0 ONLINE 0 0 0 c6t6d0 ONLINE 0 0 0 c6t7d0 ONLINE 0 0 0 c6t8d0 ONLINE 0 0 0 c6t9d0 ONLINE 0 0 0 c6t10d0 ONLINE 0 0 0 c6t11d0 ONLINE 0 0 0 errors: No known data errors # zpool iostat capacity operationsbandwidth pool used avail read write read write -- - - - - - - rpool 37.8G 890G 3 2 94.7K 17.4K tank2.03T 7.03T112 0 4.62M906 -- - - - - - - # zfs list NAME USED AVAIL REFER MOUNTPOINT rpool 39.8G 874G72K /rpool rpool/ROOT35.7G 874G18K legacy rpool/ROOT/opensolaris35.6G 874G 35.3G / rpool/ROOT/opensolaris-1 89.9M 874G 2.47G /tmp/tmp8CN5TR rpool/dump2.00G 874G 2.00G - rpool/export 172M 874G19K /export rpool/export/home 172M 874G21K /export/home rpool/swap2.00G 876G24K - tank 1.81T 6.17T 32.2K /tank tank/data 1.81T 6.17T 1.77T /data tank/public-share 34.9K 6.17T 34.9K /public-share Might also be helpful to provide the version of Opensolaris you're on, as well as the zfs version. --Tim ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] Forensics related ZFS questions
1. Does variable FSB block sizing extend to files larger than record size, concerning the last FSB allocated? In other words, for files larger than 128KB, that utilize more than one full recordsize FSB, will the LAST FSB allocated be 'right-sized' to fit the remaining data, or will ZFS allocate a full recordsize FSB for the last 'chunk' of the file? (This is a file slack issue re: how much will exist.) 2. Can a developer confirm that COW occurs at the FSB level (vs. sector level, for example)? In other words, when a single FSB (say 64KB file w/ recordsize=128KB) file is modified, and it's only one sector within that file that's modified, is it correct that what's copied-on-write is the entire 64KB FSB allocated to that file? (This is a data recovery issue.) NICOLE L. BEEBE, Ph.D., CISSP Assistant Professor The University of Texas at San Antonio Department of Information Systems Technology Management ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss