Re: [zfs-discuss] Intel X25-E SSD in x4500 followup
On Sat, 12 Sep 2009, Paul B. Henson wrote: > In any case, I agree with you that the firmware is buggy; however I > disagree with you as to the outcome of that bug. The drive is not > returning random garbage, it has *one* byte wrong. Other than that all of > the data seems ok, at least to my inexpert eyes. smartctl under Linux > issues a warning about that invalid byte and reports everything else ok. > Solaris on an x4500 evidentally barfs over that invalid byte and returns > garbage. On another note, my understanding is that the official Sun sold and supported SSD for the x4540 is basically just an OEM'd Intel X25-E. Did Sun install their own fixed firmware on their version of that drive, or does it have the same buggy firmware as the street version? It would be funny if you guys were shipping a drive with buggy firmware that just happens to work because the x4540 hardware doesn't trip over the one invalid byte :)... -- Paul B. Henson | (909) 979-6361 | http://www.csupomona.edu/~henson/ Operating Systems and Network Analyst | hen...@csupomona.edu California State Polytechnic University | Pomona CA 91768 ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Intel X25-E SSD in x4500 followup
On Sat, 12 Sep 2009, Eric Schrock wrote: > Also, were you ever able to get this disk behind a SAS transport (X4540, > J4400, J4500, etc)? It would be interesting to see how hardware SATL > deals with this invalid data. Output from 'smartctl -d sat' and > 'smartctl -d scsi' on such a system would show both the ATA data and the > translated SCSI data. My guess is that it just gives up at the first > invalid version record, something we should probably be doing. Phil Steinbachs gave you some data from an X25-E in a J4400 attached to an X4240 via an LSI 1068E based HBA, as well as one in one of the X4240's SAS slots connected to the internal Adaptec RAID controller: http://mail.opensolaris.org/pipermail/fm-discuss/2009-June/000432.html and: http://mail.opensolaris.org/pipermail/fm-discuss/2009-June/000435.html Your last email on the subject was: http://mail.opensolaris.org/pipermail/fm-discuss/2009-June/000447.html in which you said: "The primary thing is that this drive is completely busted - it's reporting totally invalid data in response to the ATA READ EXT LOG command for log 0x07 (Extended SMART self-test log). The spec defines that byte 0 must be 0x1 and that byte 1 is reserved." Phil might still be in a position to run smartctl on the drives if you're still interested in the data. I guess this is why you're now saying the drive is returning invalid data, I had forgotten the details, that was almost three months ago. In any case, I agree with you that the firmware is buggy; however I disagree with you as to the outcome of that bug. The drive is not returning random garbage, it has *one* byte wrong. Other than that all of the data seems ok, at least to my inexpert eyes. smartctl under Linux issues a warning about that invalid byte and reports everything else ok. Solaris on an x4500 evidentally barfs over that invalid byte and returns garbage. Overall, I think the Linux approach seems more useful. Be strict in what you generate, and lenient in what you accept ;), or something like that. As I already said, it would be really really nice if the Solaris driver could be fixed to be a little more forgiving and deal better with the drive, but I've got no expectation that it should be done. But it could be :). Thanks again for your help. I apologize if I've been a bit antagonistic, I tend to go "dog with a bone" when I start debating something. -- Paul B. Henson | (909) 979-6361 | http://www.csupomona.edu/~henson/ Operating Systems and Network Analyst | hen...@csupomona.edu California State Polytechnic University | Pomona CA 91768 ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS Export, Import = Windows sees wrong groups in ACLs
On Fri, Sep 11, 2009 at 8:33 PM, Owen Davies wrote: > I tried editing the /etc/group file to swap the GIDs but this didn't seem to > have the effect I wanted. Now, when I view the ACLs with an ls -V from the > OSOL side I see that the Parents group has full permissions but from the > Windows side these folders show that HOMESERVER\Kids group has full control. I think the CIFS password and group files are in /var/smb/smbpasswd and /var/smb/smbgroup.db . The latter is a SQLite 2 database that you can view with /lib/svc/bin/sqlite -G -- Brandon High : bh...@freaks.com If violence doesn't solve your problem, you're not using enough of it. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Intel X25-E SSD in x4500 followup
On Sat, 12 Sep 2009, Eric Schrock wrote: > Your statement that it is "just fine" is false: I didn't say it worked "perfectly", I said it worked "fine". Yes, it gave a *warning* that the "SMART Selective Self-Test Log Data Structure Revision Number" was 0 instead of 1, **however** other than that warning the data smartctl returned from the drive appeared correct. Results from the virgin drive: SMART Self-test log structure revision number 1 No self-tests have been logged. [To run self-tests, use: smartctl -t] Results after manually initiating self tests: SMART Self-test log structure revision number 1 Num Test_DescriptionStatus Remaining LifeTime(hours) LBA_of_first_error # 1 Extended offlineCompleted without error 00%68 - # 2 Short offline Completed without error 00%68 - The exact same drive in the x4500 running your test program to check self-test results: self-test-failure = (embedded nvlist) nvlist version: 0 result-code = 0x4 timestamp = 0x48a5 segment = 0x0 address = 0xa548a548a548 (end self-test-failure) There's definitely invalid data all right, but it's **not** originating from the drive. For that matter, the warning is about the "SMART Selective Self-Test Log Data Structure Revision Number", not the "SMART Self-test log structure revision number" -- which is correctly version 1. > Like I said, there are ways we could tighten up the FMA code to better > handle bad data before going off the rails - most likely smartctl gives > up when it sees this invalid record, while we (via SATL) keep going. > But any way you slice it, the drive is returning invalid data. The drive is not returning invalid data in a Linux box running smartctl. Other than a *warning* about the wrong revision of a data structure for a different self test, the drive seems to work just fine. I really appreciated the help you provided with figuring out what was going on with this drive in an x4500 under Solaris. I understand there's no obligation on anybody's part to make this unsupported drive work. However, given it does work correctly (at least in regards to returning smart self-test logs) under Linux, I don't see why it could not work correctly under Solaris. If it doesn't get fixed, it doesn't get fixed, but I don't understand why you're saying the drive is returning invalid data when the evidence does not support that conclusion. -- Paul B. Henson | (909) 979-6361 | http://www.csupomona.edu/~henson/ Operating Systems and Network Analyst | hen...@csupomona.edu California State Polytechnic University | Pomona CA 91768 ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Intel X25-E SSD in x4500 followup
Carson Gaspar wrote: Except you replied to me, not to the person who has SSDs. I have dead standard hard disks, and the mpt driver is just not happy. After applying 141737-04 to my Sol 10 system, things improved greatly, and the constant bus resets went away. After upgrading to OpenSolaris 6/09 things went back to being crappy. Updating to b118 did not help. And for the curious, here are one week of uniq'd log messages I receive when I'm having problems: Log info 0x31110b00 received for target 1. Log info 0x3113 received for target 0. Log info 0x3113 received for target 1. Log info 0x3114 received for target 0. Log info 0x3114 received for target 1. Log info 0x3114 received for target 3. mpt_handle_event_sync: IOCStatus=0x8000, IOCLogInfo=0x31110b00 mpt_handle_event: IOCStatus=0x8000, IOCLogInfo=0x31110b00 All disks are identical. An example iostat -nE output (note the 93 transport errors...): c7t1d0 Soft Errors: 0 Hard Errors: 6 Transport Errors: 93 Vendor: ATA Product: HDS725050KLA360 Revision: A10C Serial No: Size: 500.11GB <500107861504 bytes> Media Error: 0 Device Not Ready: 0 No Device: 6 Recoverable: 0 Illegal Request: 0 Predictive Failure Analysis: 0 -- Carson ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Intel X25-E SSD in x4500 followup
James C. McPherson wrote: On Thu, 10 Sep 2009 12:31:11 -0700 Carson Gaspar wrote: Alex Li wrote: We finally resolved this issue by change LSI driver. For details, please refer to here http://enginesmith.wordpress.com/2009/08/28/ssd-faults-finally-resolved/ Anyone from Sun have any knowledge of when the open source mpt driver will be less broken? Things improved greatly for me re: bus resets with a recent Sol 10 patch, but after my upgrade to OpenSolaris, they're back with a vengeance. An update to b118 didn't improve things, and I dare not go to anything more recent until the ZFS bug fixes hit the dev repo. From reading your blog post, it appears that mpt and fma were trying really hard to tell you that your SSD was misbehaving, and therefore you should do something about it. Turning _off_ disk fma and then totally replacing the driver with one that doesn't support fma were definitely not the recommended actions! Given the rest of this thread, I'm really keen to see (as somebody who works on mpt(7d)) how your system behaves with fixed SSD firmware, using mpt(7d) and with disk fma turned on again. After that, let's talk about "broken" drivers. Except you replied to me, not to the person who has SSDs. I have dead standard hard disks, and the mpt driver is just not happy. After applying 141737-04 to my Sol 10 system, things improved greatly, and the constant bus resets went away. After upgrading to OpenSolaris 6/09 things went back to being crappy. Updating to b118 did not help. -- Carson ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS Export, Import = Windows sees wrong groups in ACLs
> How are the parent and kids defined in the /etc/passwd file? These two are parents (names changed) : Dad:x:101:10:Dad:/export/home/Dad:/bin/bash Mom:x:102:1::/home/Mom:/bin/sh and these are the kids: Kid_a:x:103:1::/home/Kid_a:/bin/sh Kid_b:x:104:1::/home/Kid_b:/bin/sh Kid_c:x:105:1::/home/Kid_c:/bin/sh You didn't ask, but here is what the groups look like in the /etc/group file: kids::101: parents::102: family::103: > What do the ACLs look like? The ACL for my music folder, for example, is: dr-xr-xr-x+246 root root 246 Aug 26 00:16 music everyone@:r-x---a-R-c--s:fd-:allow group:kids:rwxpdDaARWcCos:fd-:allow When I went in and edited the /etc/group file so parents were GID 101 and kids were GID 102, OSOL happlily reported the ACL as: dr-xr-xr-x+246 root root 246 Aug 26 00:16 music everyone@:r-x---a-R-c--s:fd-:allow group:parents:rwxpdDaARWcCos:fd-:allow but Windows continued to report that the kids had permissions. Having read a bit more I know ZFS stores the full ACL with SID. This must then get mapped, somehow, to UNIX UIDs and GIDs and mapped a second time to CIFS users or groups. The experiment above shows that the two mappings seem to be independant; the name Windows determines for a SID does not rely at all on UNIX GIDs or SIDs. > Issues with the CIFS server are best served by asking on > cifs-discuss at opensolaris dot org So I guess what this leads me to is that you are right, I'm not really asking about ZFS or the actual ACLs and SIDs but rather how and where the mapping from ZFS SID to CIFS user/group name happens. That is obviously a topic for CIFS-Discuss. BTW, I gave up and just "chmod -R ..." to set the permissions back how I wanted them. It was still a real pain to do it that way because ZFS won't allow you to remove the last non-inherited ACL from a file or folder. Meanwhile, it will happily let Windows do just that if you are setting the permissions from there... frustrating. Thanks for the reply, Owen Davies -- 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] raidz replace issue
On 9/12/2009 10:33 PM, Mark J. Musante wrote: That could be a bug with the status output. Could you try "zdb -l" on one of the good drives and see if the label for c5t9d0 has "/old" oops, i just realized i took this thread off list. i hope you dont mind me putting it back on -- mea culpa. the data is below my sig. but we may not need it.. > appended? If so, you may be able to replace the drive by using > c5t9d0s0/old as the name. # zpool replace nfspool c5t9d0s0/old c5t9d0 cannot replace c5t9d0s0/old with c5t9d0: no such device in pool # zpool replace nfspool c5t9d0s0/o c5t9d0 # hey now!! # sleep 600 # zpool status [...] scrub: resilver in progress, 0.12% done, 108h42m to go [...] raidz1 DEGRADED 0 0 0 c3t8d0ONLINE 0 0 0 c5t8d0ONLINE 0 0 0 c3t9d0ONLINE 0 0 0 replacing DEGRADED 0 0 0 c5t9d0s0/o UNAVAIL 0 0 0 cannot open c5t9d0 ONLINE 0 0 0 woohoo! i've never had to use either "s0" or "s0/o", but hey, i'm happy. Thanks for your help. -- Jeremy Kister http://jeremy.kister.net./ ## on a disk that's online in a raidz1: # zdb -l /dev/dsk/c3t8d0s0 LABEL 0 version=4 name='nfspool' state=0 txg=13112019 pool_guid=16867309821638598147 top_guid=16762401194364239721 guid=4372736765277861814 vdev_tree type='raidz' id=9 guid=16762401194364239721 nparity=1 metaslab_array=327 metaslab_shift=31 ashift=9 asize=1199947382784 children[0] type='disk' id=0 guid=4372736765277861814 path='/dev/dsk/c3t8d0s0' devid='id1,s...@scompaq__bd3008856c__tp1012070593/a' whole_disk=1 DTL=429 children[1] type='disk' id=1 guid=246503143867597614 path='/dev/dsk/c5t8d0s0' devid='id1,s...@x0e1100eb0f79/a' whole_disk=1 DTL=428 children[2] type='disk' id=2 guid=12776584137217099681 path='/dev/dsk/c3t9d0s0' devid='id1,s...@x0e1100eb0f0e/a' whole_disk=1 DTL=427 children[3] type='disk' id=3 guid=10802333971928443637 path='/dev/dsk/c5t9d0s0/old' whole_disk=1 DTL=4722 LABEL 1 version=4 name='nfspool' state=0 txg=13112019 pool_guid=16867309821638598147 top_guid=16762401194364239721 guid=4372736765277861814 vdev_tree type='raidz' id=9 guid=16762401194364239721 nparity=1 metaslab_array=327 metaslab_shift=31 ashift=9 asize=1199947382784 children[0] type='disk' id=0 guid=4372736765277861814 path='/dev/dsk/c3t8d0s0' devid='id1,s...@scompaq__bd3008856c__tp1012070593/a' whole_disk=1 DTL=429 children[1] type='disk' id=1 guid=246503143867597614 path='/dev/dsk/c5t8d0s0' devid='id1,s...@x0e1100eb0f79/a' whole_disk=1 DTL=428 children[2] type='disk' id=2 guid=12776584137217099681 path='/dev/dsk/c3t9d0s0' devid='id1,s...@x0e1100eb0f0e/a' whole_disk=1 DTL=427 children[3] type='disk' id=3 guid=10802333971928443637 path='/dev/dsk/c5t9d0s0/old' whole_disk=1 DTL=4722 LABEL 2 version=4 name='nfspool' state=0 txg=13112019 pool_guid=16867309821638598147 top_guid=16762401194364239721 guid=4372736765277861814 vdev_tree type='raidz' id=9 guid=16762401194364239721 nparity=1 metaslab_array=327 metaslab_shift=31 ashift=9 asize=1199947382784 children[0] type='disk' id=0 guid=4372736765277861814 path='/dev/dsk/c3t8d0s0' devid='id1,s...@scompaq__bd3008856c__tp1012070593/a' whole_disk=1 DTL=429 children[1] type='disk' id=1
Re: [zfs-discuss] raidz replace issue
On 9/12/2009 9:41 PM, Mark J Musante wrote: The device is listed with s0; did you try using c5t9d0s0 as the name? I didn't -- I never used s0 in the config setting up the zpool -- it changed to s0 after reboot. but in either case, it's a good thought: # zpool replace nfspool c5t9d0s0 c5t9d0 cannot replace c5t9d0s0 with c5t9d0: no such device in pool # suex zpool replace nfspool c5t9d0s0 c5t9d0s0 cannot replace c5t9d0s0 with c5t9d0s0: no such device in pool but no luck. FYI, there are many more disks than what i showed in my previous example, but i don't think it was relevant to include them all in the email to the list. they're all working fine and are just more raidz1s. but i'll surely post the entire output of zpool status if anyone wants. -- Jeremy Kister http://jeremy.kister.net./ ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Intel X25-E SSD in x4500 followup
On Thu, 10 Sep 2009 12:31:11 -0700 Carson Gaspar wrote: > Alex Li wrote: > > We finally resolved this issue by change LSI driver. For details, please > > refer to here > > http://enginesmith.wordpress.com/2009/08/28/ssd-faults-finally-resolved/ > > Anyone from Sun have any knowledge of when the open source mpt driver will be > less broken? Things improved greatly for me re: bus resets with a recent Sol > 10 > patch, but after my upgrade to OpenSolaris, they're back with a vengeance. An > update to b118 didn't improve things, and I dare not go to anything more > recent > until the ZFS bug fixes hit the dev repo. >From reading your blog post, it appears that mpt and fma were trying really hard to tell you that your SSD was misbehaving, and therefore you should do something about it. Turning _off_ disk fma and then totally replacing the driver with one that doesn't support fma were definitely not the recommended actions! Given the rest of this thread, I'm really keen to see (as somebody who works on mpt(7d)) how your system behaves with fixed SSD firmware, using mpt(7d) and with disk fma turned on again. After that, let's talk about "broken" drivers. 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] ZFS Export, Import = Windows sees wrong groups in ACLs
Owen Davies wrote: I had a OpenSolaris server running basically as a fileserver for all my windows machines. The CIFS server was running in WORKGROUP mode. I had several users defined on the server to match my windows users. I had these users in a few groups (the most important being Parents and Kids). For various reasons I decided to blow away the OS and install fresh. First I exported the zpool holding my shared files. I then rebuilt the server, created the same users, granted the same group memberships and finally imported my zpool. Unfortunately, the GIDs for the Parents and Kids groups must have been swapped when I rebuilt because I found that the Kids had full control of everything and Parents had read only access. What is the right way to correct this? I tried editing the /etc/group file to swap the GIDs but this didn't seem to have the effect I wanted. Now, when I view the ACLs with an ls -V from the OSOL side I see that the Parents group has full permissions but from the Windows side these folders show that HOMESERVER\Kids group has full control. I know ZFS keeps the Windows style SIDs in the ACLs across export/import but how am I supposed to get the users/groups I created during the rebuild to match up with the old SIDs? Where is Windows getting a different idea of who that group is that OSOL has? There is no Active Directory or other LDAP which could have maintained a definition of those SIDs over the rebuild. Thanks, Owen Davies How are the parent and kids defined in the /etc/passwd file? What do the ACLs look like? Issues with the CIFS server are best served by asking on cifs-disc...@opensolaris.org -Mark ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] raidz replace issue
[sorry for the cross post to solarisx86] One of my disks died that i had in a raidz configuration on a Sun V40z with Solaris 10u5. I took the bad disk out, replaced the disk, and issued 'zpool replace pool c5t9d0'. the resilver process started, and before it was done i rebooted the system. now, the raidz is all upset: # zpool status pool: pool state: DEGRADED status: One or more devices could not be opened. Sufficient replicas exist for the pool to continue functioning in a degraded state. action: Attach the missing device and online it using 'zpool online'. see: http://www.sun.com/msg/ZFS-8000-D3 scrub: resilver completed with 0 errors on Sat Sep 12 17:19:57 2009 config: NAMESTATE READ WRITE CKSUM nfspool DEGRADED 0 0 0 raidz1ONLINE 0 0 0 c3t4d0 ONLINE 0 0 0 c5t4d0 ONLINE 0 0 0 c3t5d0 ONLINE 0 0 0 c5t5d0 ONLINE 0 0 0 raidz1DEGRADED 0 0 0 c3t8d0 ONLINE 0 0 0 c5t8d0 ONLINE 0 0 0 c3t9d0 ONLINE 0 0 0 c5t9d0s0/o UNAVAIL 0 0 0 cannot open raidz1ONLINE 0 0 0 c3t10d0 ONLINE 0 0 0 c5t10d0 ONLINE 0 0 0 c3t11d0 ONLINE 0 0 0 c5t11d0 ONLINE 0 0 0 spares c3t15d0 AVAIL c3t14d0 AVAIL c5t14d0 AVAIL # zpool replace nfspool c5t9d0 c5t9d0 cannot replace c5t9d0 with c5t9d0: no such device in pool # suex zpool replace nfspool c5t90d0 c5t14d0 cannot replace c5t9d0 with c5t14d0: no such device in pool Any clues on what to do here ? -- Jeremy Kister http://jeremy.kister.net./ -- Jeremy Kister http://jeremy.kister.net./ ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] alternative hardware configurations for zfs
On Sat, Sep 12, 2009 at 10:17 AM, Damjan Perenic < damjan.pere...@guest.arnes.si> wrote: > On Sat, Sep 12, 2009 at 7:25 AM, Tim Cook wrote: > > > > > > On Fri, Sep 11, 2009 at 4:46 PM, Chris Du wrote: > >> > >> You can optimize for better IOPS or for transfer speed. NS2 SATA and SAS > >> share most of the design, but they are still different, cache, > interface, > >> firmware are all different. > > > > And I'm asking you to provide a factual basis for the interface playing > any > > role in IOPS. I know for a fact it has nothing to do with error recovery > or > > command queue. > > > > Regardless, I've never seen either one provide any significant change in > > IOPS. I feel fairly confident stating that within the storage industry > > there's a pretty well known range of IOPS provided for 7200, 10K, and 15K > > drives respectively, regardless of interface. You appear to be saying > this > > isn't the case, so I'd like to know what data you're using as a reference > > point. > > I shopped for 1TB 7200rpm drives recently and I noticed Seagate > Barracude ES.2 has 1TB version with SATA and SAS interface. > > In their datasheet at > http://www.seagate.com/www/en-us/products/servers/barracuda_es/ and > product overview they claim following: > > --- > Choose SAS for the seamless Tier 2 enterprise experience, with > improved data integrity and a 135 percent average performance > boost over SATA. SAS also reduces integration complexity and > optimizes system performance for rich media, reference data > storage and enterprise backup applications. > --- > With a choice of either SATA or SAS > interfaces, the Barracuda ES.2 drive > utilizes perpendicular recording technology > to deliver the industry’s highest-capacity > 4-platter drive. SAS delivers up to a 38 > percent IOPS/watt improvement over > SATA. > --- > > And in Product overview: > --- > • Full internal IOEDC/IOECC* data integrity protection on SAS models > • Dual-ported, multi-initiator SAS provides full-duplex compatibility > and a 135 percent average** performance improvement over SATA. > > *IOEDC/IOECC on SATA (writes only), IOEDC/IOECC on SAS (both reads and > writes) > **Averaged from random/sequential, read/write activities with write cache > off > -- > > I admit I have no clue why SAS version should be/is faster. I just > pass on things I found out. But I am interested in opinion if there is > any substance in this marketing material. > > Kind regards, > Damjan > The two *'s leave much room to be desired. Averaged? How about some real numbers with testing methodology. I'm not at all surprised they claim 2x performance from the drive they charge twice as much for. On the flip side, according to storage review, the SATA version trumps the SAS version in pretty much everything but throughput (which is still negligible). http://www.storagereview.com/php/benchmark/suite_v4.php?typeID=10&testbedID=4&osID=6&raidconfigID=1&numDrives=1&devID_0=354&devID_1=362&devCnt=2 --Tim ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Problem with snv_122 Zpool issue
oh okay! But I still don't understand why is my zpool acting like this? What kind of error could this be then? I can read/write to the pool but it's going extremely slow. All my disc are fine! I'm sure about that! When I upgraded to 122 I didn't notice this problem until I rebooted after 5 days. Hope someone knows? -- 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] when will zfs have deduplication ?
tranceash wrote: Zfs will have deduplication in summer 2009 was the news ? But there seems to be no news when will it have this feature??? http://www.codestrom.com/wandering/2009/09/faq-zfs-deduplication.html ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Problem with snv_122 Zpool issue
You shouldn't hit the Raid-Z issue because it only happens with an odd number of disks. -- 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] Intel X25-E SSD in x4500 followup
Also, were you ever able to get this disk behind a SAS transport (X4540, J4400, J4500, etc)? It would be interesting to see how hardware SATL deals with this invalid data. Output from 'smartctl -d sat' and 'smartctl -d scsi' on such a system would show both the ATA data and the translated SCSI data. My guess is that it just gives up at the first invalid version record, something we should probably be doing. - Eric -- Eric Schrock, Fishworkshttp://blogs.sun.com/eschrock ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Intel X25-E SSD in x4500 followup
On Sep 12, 2009, at 12:00 AM, Paul B. Henson wrote: Well, I won't claim the drive firmware is completely innocent, but as evidenced in http://mail.opensolaris.org/pipermail/fm-discuss/2009-June/ 000436.html smartctl on a Linux box seems to work just fine. The exact same model drive also works just fine in an x4540. So I think the assertion that the drive returns random data is demonstrably false. Your statement that it is "just fine" is false: --- SMART Selective Self-Test Log Data Structure Revision Number (0) should be 1 SMART Selective self-test log data structure revision number 0 Warning: ATA Specification requires selective self-test log data structure revision number = 1 --- Like I said, there are ways we could tighten up the FMA code to better handle bad data before going off the rails - most likely smartctl gives up when it sees this invalid record, while we (via SATL) keep going. But any way you slice it, the drive is returning invalid data. - Eric -- Eric Schrock, Fishworkshttp://blogs.sun.com/eschrock ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] NFS export issue
On Sat, 12 Sep 2009, Thomas Burgess wrote: This is because with ZFS the directories aren't REALLY there. You need to either use NFSv4 or you need to export each ZFS filesystem independently It should be sufficient to use an appropriate automount rule on the client so that the "subordinate" filesystem gets mounted. That is what I do here. The problem is usually the mount from the client rather than the NFS export from the server. For example, under OS X Leopard this is what I do to access home directories exported under "freddy:/home": % cat /etc/auto_home # # Automounter map for /home # #+auto_home # Use directory service * freddy:/home/& And for FreeBSD I use this gobbly-gook: % cat /etc/home.map * type:=nfs;rhost:=freddy;rfs:=/home/${key};fs:=${autodir}/${rhost}${rfs};opts:=rw,grpid,resvport,vers=3,proto=tcp,nosuid,nodev Bob -- Bob Friesenhahn bfrie...@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer,http://www.GraphicsMagick.org/ ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] alternative hardware configurations for zfs
On Sat, Sep 12, 2009 at 7:25 AM, Tim Cook wrote: > > > On Fri, Sep 11, 2009 at 4:46 PM, Chris Du wrote: >> >> You can optimize for better IOPS or for transfer speed. NS2 SATA and SAS >> share most of the design, but they are still different, cache, interface, >> firmware are all different. > > And I'm asking you to provide a factual basis for the interface playing any > role in IOPS. I know for a fact it has nothing to do with error recovery or > command queue. > > Regardless, I've never seen either one provide any significant change in > IOPS. I feel fairly confident stating that within the storage industry > there's a pretty well known range of IOPS provided for 7200, 10K, and 15K > drives respectively, regardless of interface. You appear to be saying this > isn't the case, so I'd like to know what data you're using as a reference > point. I shopped for 1TB 7200rpm drives recently and I noticed Seagate Barracude ES.2 has 1TB version with SATA and SAS interface. In their datasheet at http://www.seagate.com/www/en-us/products/servers/barracuda_es/ and product overview they claim following: --- Choose SAS for the seamless Tier 2 enterprise experience, with improved data integrity and a 135 percent average performance boost over SATA. SAS also reduces integration complexity and optimizes system performance for rich media, reference data storage and enterprise backup applications. --- With a choice of either SATA or SAS interfaces, the Barracuda ES.2 drive utilizes perpendicular recording technology to deliver the industry’s highest-capacity 4-platter drive. SAS delivers up to a 38 percent IOPS/watt improvement over SATA. --- And in Product overview: --- • Full internal IOEDC/IOECC* data integrity protection on SAS models • Dual-ported, multi-initiator SAS provides full-duplex compatibility and a 135 percent average** performance improvement over SATA. *IOEDC/IOECC on SATA (writes only), IOEDC/IOECC on SAS (both reads and writes) **Averaged from random/sequential, read/write activities with write cache off -- I admit I have no clue why SAS version should be/is faster. I just pass on things I found out. But I am interested in opinion if there is any substance in this marketing material. Kind regards, Damjan ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Problem with snv_122 Zpool issue
Do you thing that this is a bug? If it is a bug, its okay for me. I can wait for future releases. But if this is happening only for me, then I really need help to solve this problem. -- 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] Problem with snv_122 Zpool issue
On Sat, 12 Sep 2009 07:38:43 PDT Hamed wrote: > Please help me. I really need help. I did a stupid thing i know. Afaik help does not exist in this case other than making a full backup / restore. There is no return to former zfs versions possible. -- Dick Hoogendijk -- PGP/GnuPG key: 01D2433D + http://nagual.nl/ | SunOS 10u7 5/09 | OpenSolaris 2010.02 b122 + All that's really worth doing is what we do for others (Lewis Carrol) ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] NFS export issue
This is because with ZFS the directories aren't REALLY there. You need to either use NFSv4 or you need to export each ZFS filesystem independently On Fri, Sep 11, 2009 at 4:54 PM, Thomas Uebermeier wrote: > Hello, > > I have a ZFS filesystem structure, which is basically like this: > > /foo > /foo/bar > /foo/baz > > all are from one pool and /foo does only contain the other > directories/mounts (no other files) > > When I try to export /foo via dfstab, I can see the directories bar and > baz, > but these are empty. > Can I only export via NFS each subdir on its own ? > > thanks in advance. > > Thomas > ___ > 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] Problem with snv_122 Zpool issue
Hi everyone! I did a huge mistake by upgrading my zpool from build 118 to 122. I didn't know about the checksum error. The strange thing here is that I don't get any error at all. My zpool is working very slowly, everything work fine beside the speed. It goes between 900 kb/s to 2 MB/s and accessing files/folders takes very long time. If I start a scrub it takes over 1 hours even after 3 hours it still grows. My zpool consist of 6 hard-drive with 2 TB disc using raidz2. I have export/import my zpool. everything goes fine. My problem here is the speed. I can't rollback because i also upgraded my zpool from version 14 to 16 :( now i cant go back to older builds. :( Please help me. I really need help. I did a stupid thing i know. -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] when will zfs have deduplication ?
Zfs will have deduplication in summer 2009 was the news ? But there seems to be no news when will it have this feature??? -- 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] Pulsing write performance
im playing around with a home raidz2 install and i can see this pulsing as well. The only difference is i have 6 ext usb drives with activity lights on them so i can see whats actually being written to the disk and when :) What i see is about 8 second pauses while data is being sent over the network into what appears to be some sort of in memory cache. Then the cache is flushed to disk and the drives all spring into life, and the network activity dropps down to zero. after a few seconds writing, the drives stop and the whole process begins again. Somethings funny is going on there... -- 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] sync replication easy way?
On Sep 11, 2009, at 13:40, Maurice Volaski wrote: At 8:25 PM +0300 9/11/09, Markus Kovero wrote: I believe failover is best to be done manually just to be sure active node is really dead before importing it on another node, otherwise there could be serious issues I think. I believe there are many users of Linux-HA, aka heartbeat, who do failover automatically on Linux systems. You can configure a stonith device to "shoot the other node in the head." I had heartbeat running on OpenSolaris, though I never tested failover. Did you get decent performance when you tested? Solaris Cluster is a free download I believe (even in commercial production): http://www.sun.com/software/solaris/cluster/ For OpenSolaris there's also OpenHA: http://www.opensolaris.com/learn/features/availability/ http://opensolaris.org/os/community/ha-clusters/ Need to cut a cheque for support though (for both/either). ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] Q about incremental zfs send recv
Probably a dumb (but basic) question about incremental zfs backups. After reading docs I'm still nnot sure, so I ask here. # zfs snapshot -r rpool/ROOT/b...@0901 # zfs send rpool/ROOT/b...@0901 | zfs recv -Fdu tank # zfs snapshot -r rpool/ROOT/b...@0902 # zfs send -i rpool/ROOT/b...@0901 rpool/ROOT/b...@0902 | zfs recv -vF # tank What I'd like to see confirmed is that the incremental backup is received in the -same- filesystem as the originally backup up one ("tank") -- Dick Hoogendijk -- PGP/GnuPG key: 01D2433D + http://nagual.nl/ | SunOS 10u7 5/09 | OpenSolaris 2010.02 b122 + All that's really worth doing is what we do for others (Lewis Carrol) ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] zpool hanging after I/O error (usb) on all mirror components
Hi, yesterday, my backup zpool on two usb drives failed for USB errors (I don't know if connecting my iPhone plays a role) while scrubbing the pool. This lead to all I/O on the zpool hanging, including df, zpool and zfs commands. init 6 would also hang due to bootadm hanging: process id 16329 is /sbin/bootadm -ea update_all, parent process is 1 [...] stack trace is: unix: _resume_from_idle+0xfa () unix: resume (?) unix: swtch+0x147 () genunix: cv_wait+0x61 (0xff0206e199e0,0xff0206e199d8) zfs: zio_wait+0x5d (0xff0206e19720) zfs: dbuf_read+0x1e8 (0xff01f62d3070,0,2) zfs: dmu_buf_hold+0x96 (0xff02302e51e8,3,0,0,0xff0008a66538) zfs: zap_lockdir+0x67 (0xff02302e51e8,3,0,1,1,B_FALSE,0xff0008a665c8) zfs: zap_lookup_norm+0x55 (0xff02302e51e8,3,0xff0008a66a70,8,1, 0xff0008a66778,MT_EXACT,0,0,0) [...] At this point, I didn't know of a better way than to hard reboot the machine. This happened on snv_111 running as an xvm dom0. My question is if anyone is interested in analyzing this. I'll provide some detail here and in case anyone is in interested, I could provide crash and core-dumps. Thank you, Nils -- Here's an excerpt from /var/adm/messages: Sep 11 22:01:48 haggis usba: [ID 912658 kern.info] USB 2.0 device (usb5ac,1292) operating at hi speed (USB 2.x) on USB 2.0 root hub: dev...@5, usb_mid8 at bus add ress 2 Sep 11 22:01:48 haggis usba: [ID 349649 kern.info] Apple Inc. iPhone 05a9686f1c32714410c9c14f94f800dbeaa830dd Sep 11 22:01:48 haggis genunix: [ID 936769 kern.info] usb_mid8 is /p...@0,0/pci103c,3...@1d,7/dev...@5 Sep 11 22:01:48 haggis genunix: [ID 408114 kern.info] /p...@0,0/pci103c,3...@1d,7/dev...@5 (usb_mid8) online Sep 11 22:01:48 haggis usba: [ID 349649 kern.info] usba:no driver found for interface 0 (nodename: 'image') of Apple Inc. iPhone 05a9686f1c32714410c9c14f9 4f800dbeaa830dd Sep 11 22:01:48 haggis ata: [ID 776753 kern.info] ata: simplex controller. DMA on channel 0 PIO on channel 1 Sep 11 22:01:48 haggis unix: [ID 954099 kern.info] NOTICE: IRQ18 is being shared by drivers with different interrupt levels. Sep 11 22:01:48 haggis This may result in reduced system performance. Sep 11 22:01:48 haggis xpv_psm: [ID 805372 kern.info] xVM_psm: ide (ata) instance 3 irq 0xf vector 0x98 ioapic 0x1 intin 0xf is bound to cpu 0 Sep 11 22:01:48 haggis unix: [ID 954099 kern.info] NOTICE: IRQ18 is being shared by drivers with different interrupt levels. Sep 11 22:01:48 haggis This may result in reduced system performance. Sep 11 22:01:49 haggis ata: [ID 776753 kern.info] ata: simplex controller. DMA on channel 0 PIO on channel 1 Sep 11 22:01:49 haggis unix: [ID 954099 kern.info] NOTICE: IRQ18 is being shared by drivers with different interrupt levels. Sep 11 22:01:49 haggis This may result in reduced system performance. Sep 11 22:01:49 haggis unix: [ID 954099 kern.info] NOTICE: IRQ18 is being shared by drivers with different interrupt levels. Sep 11 22:01:49 haggis This may result in reduced system performance. Sep 11 22:01:49 haggis xpv_psm: [ID 805372 kern.info] xVM_psm: ide (ata) instance 3 irq 0xf vector 0x98 ioapic 0x1 intin 0xf is bound to cpu 1 Sep 11 22:02:00 haggis xntpd[8808]: [ID 898919 daemon.error] server returns a permission denied error Sep 11 22:07:00 haggis last message repeated 23 times Sep 11 22:08:00 haggis xntpd[8808]: [ID 898919 daemon.error] server returns a permission denied error Sep 11 22:09:00 haggis last message repeated 7 times Sep 11 22:09:29 haggis scsi: [ID 107833 kern.warning] WARNING: /p...@0,0/pci103c,3...@1d,7/stor...@3/d...@0,0 (sd2): Sep 11 22:09:29 haggis Command failed to complete...Device is gone Sep 11 22:09:29 haggis genunix: [ID 408114 kern.info] /p...@0,0/pci103c,3...@1d,7/dev...@5 (usb_mid8) offline Sep 11 22:09:29 haggis ata: [ID 776753 kern.info] ata: simplex controller. DMA on channel 0 PIO on channel 1 Sep 11 22:09:29 haggis unix: [ID 954099 kern.info] NOTICE: IRQ18 is being shared by drivers with different interrupt levels. Sep 11 22:09:29 haggis This may result in reduced system performance. Sep 11 22:09:29 haggis xpv_psm: [ID 805372 kern.info] xVM_psm: ide (ata) instance 3 irq 0xf vector 0x98 ioapic 0x1 intin 0xf is bound to cpu 0 Sep 11 22:09:29 haggis unix: [ID 954099 kern.info] NOTICE: IRQ18 is being shared by drivers with different interrupt levels. Sep 11 22:09:29 haggis This may result in reduced system performance. Sep 11 22:09:30 haggis ata: [ID 776753 kern.info] ata: simplex controller. DMA on channel 0 PIO on channel 1 Sep 11 22:09:30 haggis unix: [ID 954099 kern.info] NOTICE: IRQ18 is being shared by drivers with different interrupt levels. Sep 11 22:09:30 haggis This may result in reduced system performance. Sep 11 22:09:30 haggis unix: [ID 954099 kern.info] NOTICE: IRQ18 is being shared by drivers with different interrupt levels. Sep 11 22:09:30 haggis This may result in reduced system performance. Sep 11 22:09:30 haggis xpv_psm:
Re: [zfs-discuss] Raid-Z Issue
On Sep 11, 2009, at 10:41 PM, Frank Middleton wrote: On 09/11/09 03:20 PM, Brandon Mercer wrote: They are so well known that simply by asking if you were using them suggests that they suck. :) There are actually pretty hit or miss issues with all 1.5TB drives but that particular manufacturer has had a few more than others. FWIW I have a few of them in mirrored pools and they have been working flawlessly for several months now with LSI controllers. The workload is bursty - mostly MDA driven code generation and compilation of > 1M KLoC applications and they work well enough for that. Also by now probably a PetaByte of zfs send/recvs and many scrubs, never a timeout and never a checksum error. They are all rev CC1H. So your mileage may vary, as they say... I'we also been running three of them with SD17 in a raidz for about a year without any problems at all. Regards Henrik http://sparcv9.blogspot.com ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Intel X25-E SSD in x4500 followup
On Fri, 11 Sep 2009, Eric Schrock wrote: > It's clearly bad firmware - there's no bug in the sata driver. That > drive basically returns random data, and if you're unlucky that > randomness will look like a valid failure response. In the process I > found one or two things that could be tightened up with the FMA analysis, > but when your drive is returning random log data it's impossible to > actually fix the problem in software. Well, I won't claim the drive firmware is completely innocent, but as evidenced in http://mail.opensolaris.org/pipermail/fm-discuss/2009-June/000436.html smartctl on a Linux box seems to work just fine. The exact same model drive also works just fine in an x4540. So I think the assertion that the drive returns random data is demonstrably false. There's something about the SSD in an x4500 that just doesn't play nice -- it might be partially the drive firmware, it might be the SAS controller, it might be something else -- but it's *not* simply random data being returned from the drive. It would be really appreciated if that problem could be tracked down so the drive works as well SMART-wise in an x4500 as it does in a Linux box or an x4540, but I understand Sun does not certify the x4500 with SSD's so there's no expectation that would happen. But it would be really really appreciated :)... -- Paul B. Henson | (909) 979-6361 | http://www.csupomona.edu/~henson/ Operating Systems and Network Analyst | hen...@csupomona.edu California State Polytechnic University | Pomona CA 91768 ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss