September 30, 2013 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >
September 30, 2013 [00:04:42] *** lgtaube has quit IRC[00:13:46] *** jaimef has joined #openindiana[00:18:13] *** szaydel has joined #openindiana[00:18:44] *** newbie14 has joined #openindiana[00:22:42] *** newbie12 has quit IRC[00:25:26] *** Vutral has quit IRC[00:26:48] *** Vutral has joined #openindiana[00:29:06] *** Vutral has quit IRC[00:30:20] *** andy_js has quit IRC[00:31:33] *** jaimef has quit IRC[00:51:27] *** Vutral has joined #openindiana[00:59:41] *** drlou_on_tour has joined #openindiana[01:01:15] *** jaimef has joined #openindiana[01:06:32] *** lgtaube has joined #openindiana[01:07:00] *** SonikkuAmerica has quit IRC[01:13:14] *** szaydel has quit IRC[01:17:42] *** sac_ has quit IRC[01:18:06] *** sac`away has quit IRC[01:18:20] *** sac_ has joined #openindiana[01:18:26] *** sac`away has joined #openindiana[01:23:36] *** phobos_anomaly has joined #openindiana[01:28:43] *** szaydel has joined #openindiana[01:32:38] *** lgtaube has quit IRC[01:34:18] *** lgtaube has joined #openindiana[01:39:44] *** szaydel has quit IRC[01:41:41] *** lgtaube has quit IRC[01:41:48] *** Lee- has quit IRC[01:45:28] *** lgtaube has joined #openindiana[02:04:26] *** sjorge has quit IRC[02:09:57] *** sjorge has joined #openindiana[02:27:54] *** InTheWings has quit IRC[02:32:53] *** szaydel has joined #openindiana[02:37:06] *** herzen has quit IRC[02:40:57] *** drlou_on_tour has quit IRC[02:44:06] *** szaydel has quit IRC[02:51:55] *** szaydel has joined #openindiana[02:55:29] *** dubf has quit IRC[02:56:15] *** dubf has joined #openindiana[03:00:23] *** sac`away has quit IRC[03:01:35] *** sac`away has joined #openindiana[03:27:00] *** herzen has joined #openindiana[03:33:05] *** hlangeveld has quit IRC[03:39:04] *** ngchk1 has joined #openindiana[03:41:48] *** ngchk1_ has quit IRC[04:06:34] *** szaydel has quit IRC[04:13:42] *** SonikkuAmerica has joined #openindiana[04:14:30] *** SonikkuAmerica has quit IRC[04:17:01] *** SonikkuAmerica has joined #openindiana[04:18:35] *** newbie12 has joined #openindiana[04:21:30] *** newbie14 has quit IRC[04:35:27] *** joshua_ has quit IRC[04:59:20] *** SadMan has quit IRC[04:59:57] *** SadMan has joined #openindiana[05:00:40] *** rocketeer has joined #openindiana[05:05:10] *** SonikkuAmerica has quit IRC[05:08:44] *** SonikkuAmerica has joined #openindiana[05:25:08] *** gm152 has quit IRC[05:39:23] *** szaydel has joined #openindiana[05:58:18] *** Anton_ has quit IRC[06:03:48] *** Anton_ has joined #openindiana[06:16:54] *** Lee- has joined #openindiana[06:17:58] *** zerobsd has quit IRC[06:18:58] *** davenz has quit IRC[06:22:16] *** ce_cute_25 has joined #openindiana[06:22:23] *** ce_cute_25 has left #openindiana[06:23:53] *** alex1 has joined #openindiana[06:40:24] *** man_u_ has joined #openindiana[06:40:57] *** szaydel has quit IRC[06:41:07] *** wkemp_ has joined #openindiana[06:41:48] *** wonko2 has joined #openindiana[06:42:35] *** melliott_ has joined #openindiana[06:43:39] *** ashkan_ has joined #openindiana[06:43:53] *** sommerfeld_ has joined #openindiana[06:43:55] *** n2deep_ has joined #openindiana[06:43:58] *** ivan\_ has joined #openindiana[06:44:00] *** d8r has joined #openindiana[06:44:03] *** kovert1 has joined #openindiana[06:44:18] *** atrius_ has joined #openindiana[06:45:13] *** dijenerate_ has joined #openindiana[06:49:39] *** melliott has quit IRC[06:49:40] *** wkemp has quit IRC[06:49:41] *** sommerfeld has quit IRC[06:49:42] *** dijenerate has quit IRC[06:49:44] *** ivan\ has quit IRC[06:49:44] *** wonko has quit IRC[06:49:45] *** kovert has quit IRC[06:49:47] *** ashkan has quit IRC[06:49:48] *** n2deep has quit IRC[06:49:49] *** man_u has quit IRC[06:49:50] *** digidigger has quit IRC[06:49:50] *** atrius` has quit IRC[06:49:52] *** ashkan_ is now known as ashkan[06:49:52] *** atrius_ is now known as atrius`[06:49:53] *** ivan\_ is now known as ivan\[06:49:54] *** man_u_ is now known as man_u[06:54:40] *** shruti has joined #openindiana[07:08:11] *** phobos_anomaly has quit IRC[07:15:22] *** basic6_ has joined #openindiana[07:15:58] *** basic6 has quit IRC[07:40:49] *** nikolam has quit IRC[07:47:38] *** nikolam has joined #openindiana[07:48:52] *** newbie has joined #openindiana[07:49:16] *** newbie is now known as Guest83714[07:52:54] *** newbie12 has quit IRC[08:00:49] *** nikolam has quit IRC[08:03:40] *** newbie12 has joined #openindiana[08:04:15] *** nikolam has joined #openindiana[08:07:01] *** Guest83714 has quit IRC[08:07:33] *** rfm has quit IRC[08:19:43] *** Seony has joined #openindiana[08:55:36] *** tsoome has quit IRC[09:03:12] *** alcir has joined #openindiana[09:04:46] *** Seemone has joined #openindiana[09:09:00] *** |AbsyntH| has joined #openindiana[09:20:01] *** texarcana has quit IRC[09:21:47] *** tsoome has joined #openindiana[09:22:07] *** texarcana has joined #openindiana[09:22:31] *** alanc has quit IRC[09:22:49] *** alanc has joined #openindiana[09:22:49] *** ChanServ sets mode: +o alanc[09:26:48] *** alcir has quit IRC[10:18:01] *** andy_js has joined #openindiana[10:18:16] *** alcir has joined #openindiana[10:41:58] *** andy_js_ has joined #openindiana[10:44:59] *** andy_js has quit IRC[10:45:00] *** |AbsyntH| has quit IRC[10:45:03] *** andy_js_ is now known as andy_js[10:45:34] *** |AbsyntH| has joined #openindiana[10:57:34] *** Seony^Zzz has joined #openindiana[10:57:35] *** Seony has quit IRC[11:12:06] *** jihyun has quit IRC[11:15:11] *** KermitTheFragger has joined #openindiana[11:16:44] *** |AbsyntH| has left #openindiana[11:16:55] *** InTheWings has joined #openindiana[11:20:36] *** jihyun has joined #openindiana[11:27:05] *** movement has joined #openindiana[11:30:35] *** Seony^Zzz has quit IRC[11:31:13] *** nikolam has quit IRC[11:33:34] *** newbie13 has joined #openindiana[11:34:50] *** Peanut has joined #openindiana[11:36:56] <Seemone> I want to measure how much data is writte to my ZIL/L2ARC SSDs. sar doesn't help as it reports reads and writes all together. the next thing I'm thinking of is leaving a zpool iostat running. is there a "better" method?[11:37:08] *** newbie12 has quit IRC[11:39:28] *** Worsoe has joined #openindiana[11:39:35] <Peanut> Hi folks. My console is showing "Warning: /pci@0,0/pci-ide@14,1/ide@0 (ata0): timeout : abort request, target=1, lun=0. How could I find out which disk it is? My zpool status doesn't show anything being degraged.[11:40:22] <Seemone> iostat -E[11:41:06] *** rocketeer has quit IRC[11:41:45] <Seemone> actually[11:41:47] <Seemone> iostat -En|grep Error|grep -v "No D"|awk '{print "echo "$7"."$10" "$1" $(zpool status |awk '\''/pool:/{pool=$2}/d0/{print pool\" \"$1}'\''|grep "$1"|cut -d\" \" -f 1)"}'|sh|sort -n|awk '$3!=""'[11:42:00] <Seemone> outputs things like[11:42:00] <Seemone> 0.0 c1t0d0 data[11:42:01] <Seemone> 0.0 c1t1d0 data[11:42:01] <Seemone> 0.0 c1t2d0 data[11:42:25] *** rocketeer has joined #openindiana[11:42:27] <Seemone> <soft errors>.<hard errors> <disk name> <pool>[11:42:33] <Peanut> Thanks, Seemone. iostat -E shows 13 illegal request for EACH of my 4 drives in the zpool.[11:43:17] <Seemone> it also filter out "No Device" errors from optical drives[11:43:20] <Seemone> you're welcome[11:45:18] <Peanut> Apparently, "illegal request" counts increasing are a bit of a known issue with OI?[11:45:33] *** alcir has quit IRC[11:45:38] <Seemone> dunno, never had them[11:48:34] *** churnd has quit IRC[11:48:35] <Alasdairrr> illegal requests can come from smartctl if you're using that[11:48:39] *** newbie12 has joined #openindiana[11:48:58] *** churnd has joined #openindiana[11:49:04] <Alasdairrr> it issues commands to the drives that the illumos scsi layer doesn't grok[11:50:12] <Peanut> Is smartctl installed by default? My OI is stock, I haven't added anything.[11:50:32] <Peanut> I just started a 'zpool scrub' and now the count is up to 15 for each drive.[11:52:14] *** newbie13 has quit IRC[11:52:41] <Alasdairrr> Illegal requests aren't a serious issue, although they may indicate something isn't understanding something else[11:53:17] <Alasdairrr> So it could be down to shitty sas/sata/whatever drivers, sas/sata/whatever firmware, port expanders, weird harddrives, etc etc[11:53:26] <Alasdairrr> Without a full description of your setup its impossible to help[11:54:36] <Alasdairrr> A description such as "I'm using a LSI 9207-8i with the latest firmware directly attached to 4 Seagate Foo 3.5" 2TB SATA drives" would be helpful :-)[11:55:18] <Alasdairrr> If it's "I'm using some random motherboard with internal SATA to 4 brand new OCZ SATA SSDs" then I'd expect Illegal requests ;-)[11:58:30] <Peanut> Ok. Setup is a HP N40L, drives are WD 1TB Red. Trying to find out the controller now.[11:59:52] <Alasdairrr> Is it just attached to the internal SATA?[12:00:09] <Alasdairrr> It'll almost certainly be down to a shitty driver I'm afraid[12:01:00] <Peanut> Yes, 4 disks on the 4 internal sata ports, and an SSD for the OS in the CDROM bay.[12:01:18] <Alasdairrr> I imagine it'll be the Intel AHCI stuff[12:01:26] <tomww> Seemone: there are kstat variables counting the bytes written / read from a device. That and maybe a device specific variable read by smartctl telling how many blocks or bytes or whatever the SSD thinks on how to name that.[12:01:27] <Alasdairrr> Do you have a PCIe slot?[12:01:52] <Alasdairrr> You could pick up an LSI 3081E-R off ebay for very cheap[12:01:56] <Seemone> I don't have smart, silly hp smartarray[12:02:03] <Peanut> Nothing is installed in the PCI-E slots at the moment.[12:02:14] <Seemone> I'll look into kstat, thanks[12:04:25] <Peanut> Alasdairrr: that's a raid card? Wouldn't a 'simple' HBA suffice for ZFS, or do such cards simply not exist anymore?[12:05:04] <Alasdairrr> It's a very common older class of LSI card, sadly they can only do up to 2TB drives, but they're very cheap - they accept both IT and IR firmware so can be used in target HBA mode or RAID mode[12:05:50] <Alasdairrr> if you had the cash then an LSI 9207-8i is obviously a nice card :-)[12:06:03] <Alasdairrr> oh there's the 3041E-R[12:06:13] <Alasdairrr> thats a 4 port one so a bit cheaper[12:07:11] <Peanut> Ah.. but that's still almost as expensive as the N40L was.[12:07:20] <tomww> Seemone: smartctl schoul dbe availabel from the oi extra repository "sfe" if its not in the regular repo[12:08:16] <Peanut> It seems that every time I type 'zpool status', each drive gets 2 more Illegal Requests.[12:08:56] <tomww> how does zpool status print illegal requests?[12:09:03] <Seemone> tomww: I can't access the drive smart data because I have the SSD behind an HP smartarray. no jbod and no hp smartarray passthru support in smartctl for OI[12:09:11] <tomww> is you partitioning right?[12:09:20] <tomww> ah okay.[12:09:20] <Peanut> Zpool status doesn't print anything, but if I run iostat -E before and after, I can see the correlation.[12:09:46] <tomww> but kstat variables should help, they count data for every single device.[12:22:00] <Peanut> Some more digging leads me to think that the orriginal console error message was not for any of my zpool drives, but actually the SSD drive throwing an error.[12:32:15] <tomww> about illegal requests, it might be that there is a SATA command just tried but not supported by the drive[12:40:14] *** alcir has joined #openindiana[12:42:18] <Seemone> tomww: I found the right kstat stat. thanks[12:45:08] *** alcir has quit IRC[12:46:44] <tomww> "write_by" ?[12:47:52] <tomww> I would like know how much data has been written per serial number of each SSD. That way one could get and idea how long the SSD works, and as a[12:48:23] <tomww> different measure, after how many terrabytes the SSD starts to slow down (either re-org or other effekts)[12:50:25] *** alcir has joined #openindiana[12:50:28] *** InTheWings has joined #openindiana[12:51:14] <Seemone> nwritten[12:51:24] <Seemone> kstat -n sd2[12:51:41] <Seemone> while : ; do kstat -n sd2|awk '/^$/{sd2=0}$1=="name:"&&$2=="sd2"{sd2=1}sd2==1&&$1=="nwritten"{nwritten=$2; print systime()" "nwritten}';sleep 60;done|awk 'NR>1{printf "%.2f\n",($2-oldv)/($1-oldt)/1024.0/1024.0}{oldv=$2;oldt=$1}'[12:51:59] <Seemone> avg written MB/s in a 60 seconds span[12:52:24] <Seemone> I also put[12:52:25] <Seemone> * * * * * kstat -n sd2,\[ab\] -p -T u >>/root/sd2.mon[12:52:43] <Seemone> sd2,a is the ZIL sd2,b is the L2ARC[12:53:08] *** movement has quit IRC[12:53:17] <Seemone> to storicize it[12:53:21] *** movement has joined #openindiana[12:54:31] <tomww> okay, thanks. will look into this later and see how that compares to the value write_by.[12:54:46] <Seemone> smartctl is boot resistant[12:54:50] <Seemone> kstat is not[12:55:04] <tomww> preferred woudl be to map the actual serial number of the device, as the sd"instance" can change over time, but the hardware entity stays the same.[12:55:33] <Seemone> yeah.[12:55:36] <tomww> yes, you need a permanent store[12:55:46] <Seemone> I believe that it stays the same lest the hardware changes[12:56:14] <tomww> if you remove device links and path_to_inst, then it is very likely that the sd"instance" changes.[12:56:40] <Seemone> that's what I meant[12:56:45] <tomww> so the permanent store should be based on e.g. the device's serial number (at least that is what I'll try to implement in my setup)[12:57:03] <Seemone> let me know if you have a way to map them[12:57:09] <tomww> well, we have several admins around, so that is only a matter of time :)[12:57:27] <Seemone> my workload is fairly constant, I'll be happy with an estimation[12:58:24] *** alcir has quit IRC[12:58:42] <tomww> at least "iostat -E" can print the SD"instance" and Serial number for a device[12:59:24] <tomww> and parting this gets more easy, if one calls "iostat -E sd2"[12:59:32] <tomww> *parsing[13:00:19] <Seemone> fuxking hp smartarray… 2 logical drives, same serial number[13:00:35] <Seemone> I also have illegal requests...[13:02:01] <tomww> oh. if the serial number dissapears by the controler inbetween...[13:02:24][13:03:07] <Seemone> nope[13:04:24] *** newbie12 has quit IRC[13:04:38] *** newbie13 has joined #openindiana[13:04:43] *** smrt has quit IRC[13:04:50] *** newbie13 is now known as Zenos[13:05:01] *** smrt has joined #openindiana[13:15:02] <Seemone> my l2arc writes more than the my zil[13:17:40] <Seemone> not really. more or less the same atm[13:18:16] <Seemone> once the seconds SSD is online the amount of writes per SSD for the l2arc should halve[13:19:15] <Seemone> I'm leaving a few weeks in between putting them online so they don't wear out at the same time[13:23:58] <Seemone> if there is no write amplification I can expect ~24 years of life[13:24:00] <tomww> what SSD model? I user neutron for small setups, they might[13:24:16] <Seemone> samsung 840pro (I know, no capacitor)[13:24:19] <Seemone> 512GB[13:24:23] <tomww> not fall into the write-slowdown category[13:24:38] <Seemone> I partizioned them leaving 25% spare area, this should ensure a smooth latency[13:24:54] <Seemone> according to anandtech tests[13:25:04] <Seemone> 20GB ZIL (big, I know)[13:25:10] <Seemone> 337GB L2ARC[13:25:13] <Seemone> each.[13:25:21] <tomww> if you have something like munin, you could record IO completion times.[13:25:27] <Seemone> I'm writing 175GB/day or so[13:25:42] <Seemone> well my ARC has hit ratio 97-99%[13:25:53] <Seemone> the working set is fairly small[13:26:49] <Seemone> the ssd is there more for the ZIL than for the l2arc. 99% of the load is NFS[13:27:31] <Seemone> and the controller handshakes at 1.5GBps as of now[13:27:39] <Seemone> (firmware upgrade planned)[13:28:39] <Seemone> I moved form a very old amd cpu with 7 mirrored vdevs (half a TB each) to this xeon with 4 mirrored vdevs (2TB each)[13:29:12] <Seemone> so I have less spindles but I have the SSDs which I didn't have[13:29:40] <Seemone> in the old server the bottleneck was the on pool ZIL + cpu (compression is enabled)[13:34:03] *** szaydel has joined #openindiana[13:55:26] *** drlou_on_tour has joined #openindiana[14:05:47] *** WormDrink has joined #openindiana[14:16:54] *** shruti has quit IRC[14:19:17] *** Kralian has joined #openindiana[14:29:24] <mui_> hm[14:29:26] <mui_> genunix: [ID 408114 kern.info] /pci@33,0/pci8086,3c04@2/pci1000,3020@0/iport@v0 (mpt_sas3) quiesced[14:29:33] <mui_> what this quiesced means?[14:29:46] <mui_> controller that used to work on different server, resists on this one[14:43:42] *** DrLou has quit IRC[14:55:36] *** nikolam has joined #openindiana[14:56:29] *** alcir has joined #openindiana[15:03:05] *** nikolam has quit IRC[15:06:20] *** drlou_on_tour has quit IRC[15:06:22] *** robinsmidsrod has quit IRC[15:10:38] *** robinsmidsrod has joined #openindiana[15:27:17] *** nikolam has joined #openindiana[15:27:52] *** nikolam has quit IRC[15:34:06] *** DrLou has joined #openindiana[15:34:06] *** ChanServ sets mode: +o DrLou[15:37:27] *** nikolam has joined #openindiana[15:42:38] *** DerSaidin has joined #openindiana[15:46:57] *** zerobsd has joined #openindiana[15:53:39] *** alcir has quit IRC[15:56:29] *** TomH has joined #openindiana[15:59:32] <TomH> Hi All, Can anyone help with root access over NFSv4? It seems to be being squashed to nobody despite a root=@ in the zfs set sharenfs.[16:00:16] <MarcelT> try anon option[16:00:39] <Agnar_> TomH: the voodoo sits in sec= parameter. unfortunately I had no time to care for and switched to nfsv3[16:01:16] <tsoome> NFSv4 needs to have same domain, if domain wont match, all users will be mapped to nobody[16:01:47] <TomH> Hi tsoome we have the same domain defined on both the client and server, and any non root users work fine...[16:02:36] <tsoome> the root= option is set how? by using host name? does that host name resolve in server?[16:03:00] <tsoome> more importantly, does IP resolve to name?[16:03:59] <TomH> Agnar_ is sec= an option in sharenfs ?[16:04:36] <tsoome> yes it is[16:04:44] <tsoome> man share_nfs will list them all[16:04:56] *** Worsoe has quit IRC[16:05:47] *** unsound_ has quit IRC[16:06:16] <TomH> zfs set sharenfs=root=rw= at x dot x.x.x/26:x.x.x.x/26 tank/nfs and also zfs set sharenfs=rw= at x dot x.x.x/26:x.x.x.x/26,root=@x.x.x.x/26 tank/nfsp01 as well as defining the host in hosts file and directly in[16:06:50] *** mortal has joined #openindiana[16:07:47] <Wraithh> TomH, try to add sec=sys, in front[16:07:55] <Agnar_> I should migrate our solaris AI server from OI to sol11..*humm*[16:08:43] <tsoome> for some clients sec= option can do, yep[16:09:18] *** joltman has joined #openindiana[16:10:35] <TomH> Wraithh i will try that now cheers[16:11:56] *** tsoome has quit IRC[16:12:16] *** alcir has joined #openindiana[16:13:36] *** rfm has joined #openindiana[16:13:47] <TomH> zfs set sharenfs=sec=sys,root=rw= at x dot x.x.x/26:x.x.x.x/26 tank/nfs[16:14:13] <Wraithh> try root= at x dot x.x.x/26 as well[16:16:00] <Wraithh> so sharenfs=sec=sys,root= at x dot x.x.x/26,rw=@x.x.x.x/26:x.x.x.x/26[16:16:22] <TomH> Ohhh Wraithh i think thats the ticket![16:17:19] <TomH> Let me go test..[16:23:00] *** alcir has quit IRC[16:26:25] *** ianj has quit IRC[16:27:25] *** djfirefox has joined #openindiana[16:33:31] *** nikolam has quit IRC[16:35:41] <djfirefox> hey guys.[16:36:40] <djfirefox> I created a couple of iSCSI Volumes, and then only realised now I needed to leave unused space to snapshot the iSCSI LUNS, although I do have 156Gb free it seems it isn't enough to allow a snapshot[16:37:07] <djfirefox> is there any way to make an exisitning zfs Volume a sparse 'thin' volume on the fly?[16:38:59] <patdk-wk> nope[16:39:07] <djfirefox> crap[16:39:11] <patdk-wk> but I don't understand how 156gb is not enough space to make a snapshot[16:39:27] <djfirefox> can i reduce the size of a fixed disk?[16:39:44] <djfirefox> well the 156gb free is pace free due to compression[16:39:46] <patdk-wk> have no idea what that even means[16:39:59] <djfirefox> other than that I dont know, it just says 'out of space' when I try[16:40:11] <patdk-wk> so in other words, that is not really free space at all, but reserved space[16:41:19] <djfirefox> well its a 3.41TB ZFS FS with two volumes in it 1 x 1.36T and 2.05T[16:41:25] <lblume> If a dataset was created with a volsize, yes[16:41:26] <djfirefox> with lz4 compression[16:41:54] <djfirefox> and 156G free in root pool and lots of free space on the fixed Volumes themselves[16:42:08] <djfirefox> but i stupidly didn't mke them thin :([16:43:09] <lblume> datasets created with volsize require that the free space be >= to their own size to take a snapshot.[16:43:50] <djfirefox> FS is vmbkp http://pastebin.com/f7MprNYx[16:44:07] <djfirefox> WHAT!?[16:44:15] <djfirefox> >= !!![16:44:36] <lblume> Yes. I hit that issue recently with swap.[16:44:43] <djfirefox> crap crap crap[16:44:48] <djfirefox> how do I fix it[16:45:00] <djfirefox> I have an identical spare box if that helps[16:45:18] <djfirefox> but then I cant even zfs send recieve with out space[16:46:17] <lblume> I think you might be able to if you modify the stream so that the volsize is not sent over.[16:46:38] <djfirefox> so unfortunately storage snapshots aren't really possible with ZFS the likes of a storage array like emc, which can take snpshots of whatever you like and just change block track[16:47:18] <djfirefox> yes but you cant zfs send an FS without a snapshot, and I cant take one :([16:47:33] <lblume> Maybe if you use thin provisioning or zfs datasets directly instead of block volumes.[16:48:34] <djfirefox> zfs datasets?[16:48:38] <patdk-wk> the issue is, you created the dataset with reserved diskspace, so the system wants to make sure it is impossible for you to run out of diskspace to save your zvol data[16:49:38] <djfirefox> I've just tried to take a snapshot of the root pool and it worked :)[16:49:48] <djfirefox> not ideal but at least I can send the data elsewhere[16:49:55] <lblume> So practically means you need a pool at least twice as big as your biggest volsize.[16:50:21] <patdk-wk> at snapshot creation time[16:50:31] <patdk-wk> you need 2x vol size + snapshot used space[16:51:28] <djfirefox> ok so, a few questions,[16:51:35] <djfirefox> can i shrink a thich volume[16:51:39] <djfirefox> thick[16:52:10] <lblume> Shrinkage in ZFS appears to be done by the same team that worked on Duke Nukem Forever.[16:52:22] <patdk-wk> what is thick?[16:52:29] <djfirefox> opposite of thin[16:52:35] <djfirefox> ie not sparse[16:52:46] <djfirefox> created with - V not -V -s[16:52:49] <patdk-wk> well, zfs has sparse and not sparse[16:52:55] <patdk-wk> vmware has thick/thin/...[16:52:59] <patdk-wk> the terms are not compatable[16:53:00] <djfirefox> lol[16:53:00] <xenol> djfirefox: did you specify "-s" flag while creating ZVOL?[16:53:04] <djfirefox> nope[16:53:21] <xenol> djfirefox: please provide "zfs get all zvol" output[16:53:27] <djfirefox> and i extended one of them after compression with volsize[16:53:31] <xenol> zvol being iscsi lun[16:55:59] <djfirefox> http://pastebin.com/xejXAY2R[16:56:21] <patdk-wk> hmm, that works heh[16:56:35] <patdk-wk> zfs set refreservation=none zvol[16:56:43] <xenol> yes, and you have sparse zvol[16:56:55] <xenol> but be warned if you run out of space[16:59:27] <djfirefox> so if its sparse how come its taking up full allocated space in root pool?[16:59:34] <patdk-wk> it's not sparse[16:59:50] <patdk-wk> refreservation > 0, it's not sparse[17:00:09] <xenol> it will be sparse once you remove refreservation.[17:00:13] <xenol> to correct myself :)[17:00:30] <djfirefox> reservation says none??[17:00:32] <patdk-wk> I hadn't realized that was writable[17:00:41] <patdk-wk> refreservation != reservation[17:00:53] <patdk-wk> hard time reading?[17:01:46] <patdk-wk> reservation, free space, that is reserved[17:01:54] <patdk-wk> refreservation, used space, that is reserved[17:02:21] <patdk-wk> or likely a lot better way to say it[17:02:24] <xenol> djfirefox: your dataset is using 1.05T, rest is refreservation[17:02:34] <TomH> Wraithh your a legend! Cheers[17:02:40] <xenol> you can check usedby* attributes[17:02:55] <djfirefox> ooooooooh refreservation i was looking at reservation[17:03:16] *** master_o1_master has joined #openindiana[17:03:19] <djfirefox> so set refre to 0[17:03:21] <djfirefox> ?[17:03:23] <patdk-wk> none[17:03:38] <patdk-wk> man, even paste the exact command, and it does no good[17:04:12] <djfirefox> sorry didnt see command my bad[17:04:23] <patdk-wk> that set has horrible compression though[17:04:52] <djfirefox> i thought lz4 was best?[17:05:07] <djfirefox> not too heavy on cpu but good comp ratio[17:05:13] <xenol> lz4 is best for compression[17:05:29] <Agnar_> compressed things can't be well compressed again...so if you store zips, gzips, jpegs, mpegs, etc...[17:05:31] <xenol> but not every workload can be compressed[17:06:56] *** master_of_master has quit IRC[17:07:23] <xenol> e.g I am running OI file server, which stores video archives in mpeg and the compression sucks. However, setting that on vmware iSCSI pool gives me 2.02x compressratio :)[17:07:31] <djfirefox> its a backup drive on windows for veeam so i get a little better compression with lz4[17:07:53] <patdk-wk> yuk :([17:08:00] <patdk-wk> why not just let veeam use it directly?[17:08:11] <patdk-wk> veeam will just ssh in and run a small perl script[17:08:18] <patdk-wk> no need to mess with zvol's[17:08:26] <patdk-wk> and then you have direct files on the system[17:10:50] *** sommerfeld_ has quit IRC[17:11:01] *** sommerfeld has joined #openindiana[17:11:26] <djfirefox> I wanted to use iSCSI, as OI server is attached to iSCSI storage layer, veeam is a VM on ESXi, so i've attached iSCSI LUNS on OI applicance as RDMs to veeam VM[17:12:11] <patdk-wk> I'm just saying, that is overkill, and a maintance burden, as you just found out[17:12:15] <djfirefox> I then snapshot these iSCSI vols and send them over the WAN to DR site OI Storage applicance[17:12:15] *** tsoome has joined #openindiana[17:12:32] <patdk-wk> if you just had veeam access oi via ssh, it could just save them as normal files, instead of on a lun[17:15:36] <djfirefox> I appreciate that, but I need to use iSCSI, for traffic, 8 NICS connected via iSCSI , SSH would mean 1 NIC 1 session.[17:16:00] <patdk-wk> ah, your limited via gigabit[17:16:08] <djfirefox> yup yup yup unfortunately[17:17:09] <Agnar_> 802.3ad (trunking) is no option?[17:17:21] <patdk-wk> agnar, read up on what 802.3ad is[17:17:40] <djfirefox> no ciscos using netgears so can have LACP LAGS, but they didnt work well last time i tried[17:17:44] <Agnar_> patdk-wk: link trunking, isn't it?[17:17:52] <Agnar_> djfirefox: i see[17:17:59] <patdk-wk> agnar, no[17:18:14] <Agnar_> my fault then[17:18:33] <patdk-wk> it's link bonding, maybe you mean[17:18:40] <patdk-wk> link trunking would be like, vlans?[17:18:48] <Agnar_> patdk-wk: aaaaargs! :)[17:18:53] <patdk-wk> but still, it in no way has any effect[17:18:58] <Agnar_> vlans == tagging (802.1q)[17:19:15] <Agnar_> trunking is link combination :)[17:19:19] <Agnar_> except at cisco :)[17:19:29] <patdk-wk> trunking is not combining[17:19:36] <patdk-wk> trunking is vlan thing[17:19:59] <patdk-wk> link aggregation is the offical term[17:20:00] <djfirefox> cisco has etherchannel so 8 1 gb nics becomes a nice big 8Gbit trunk up and down LACP LAGS just allow 8 machines to connect at 1GB[17:20:32] <Agnar_> patdk-wk: depends on which documentation you read. if I had to do with vlans, it was always vlan-tagging[17:20:48] <patdk-wk> 802.3ad limits a single mapping to 1 connection[17:20:57] <patdk-wk> I have never seen trunking used in a 802.3ad document :)[17:21:26] <patdk-wk> if you setup 802.3ad to hash based on port number, then you get one connection per nic[17:21:32] <patdk-wk> that is the best it can do though[17:21:41] <patdk-wk> tcp does not like out of order packets at all[17:21:43] <tsoome> "LACP, otherwise known as IEEE 802.1ax Link Aggregation Control Protocol, is simply a way to dynamically build an EtherChannel."[17:22:07] <djfirefox> tsoome is on the ball ;)[17:22:16] <patdk-wk> 802.1ax != 802.3ad[17:22:21] <Agnar_> tsoome: which will enable more bandwidth with multiple connections[17:22:46] <Agnar_> .1ax is iirc not supported on solaris, eh?[17:22:50] <djfirefox> bassically without cisco, LAGS are all you got and they aren't great for high bandwidth single connections, thus enters multi-path iSCSI :)[17:23:13] *** rfm has quit IRC[17:23:36] <tsoome> EtherChannel aggregates the traffic across all the available active ports in the channel. The port is selected using a Cisco-proprietary hash algorithm, based on source or destination MAC addresses, IP addresses or TCP and UDP port numbers. The hash function gives a number between 0 and 7, and the following table shows how the 8 numbers are distributed among the 2 to 8 physical ports. In the hypothesis of real random hash algorithm, 2, 4 or[17:23:37] <tsoome> ports configurations lead to fair load-balancing, whereas other configurations lead to unfair load-balancing.[17:23:50] <xenol> Agnar_: .1ax is a LACP, which is supported by solaris[17:24:05] <tsoome> how on earth they can tell its 8G/s (with 8x1G/s links) is beyond me.[17:24:23] <xenol> tsoome: marketing[17:24:29] <tsoome> exactly[17:24:32] <Agnar_> xenol: ah, via dladm create-aggr -L - just looked up in the man[17:25:25] *** Zenos has quit IRC[17:32:22] *** ianj has joined #openindiana[17:34:18] *** joltman has quit IRC[17:35:24] *** spanglywires has joined #openindiana[17:38:30] <Wraithh> TomH, Glad I could help![17:39:14] *** gosx has joined #openindiana[17:45:17] <djfirefox> is this better http://pastebin.com/MMbURPFE[17:54:13] *** WormDrink has quit IRC[17:58:59] <xenol> djfirefox: yes, because "available" just grew from 4xx G to 1.44T[17:59:33] *** sjorge has quit IRC[17:59:51] <xenol> though turning off compression would spare some cpu cycles as there is no point in compressing already compressed files.[18:00:23] *** sjorge has joined #openindiana[18:02:24] *** sjorge has quit IRC[18:02:45] <djfirefox> surprisingly the compression is really good on the second lun[18:03:27] *** Seemone has quit IRC[18:07:54] *** WormDrink has joined #openindiana[18:14:36] *** spanglywires has quit IRC[18:17:20] *** hunter has joined #openindiana[18:23:08] *** held has quit IRC[18:23:09] *** magyar has joined #openindiana[18:23:10] *** magyar has joined #openindiana[18:32:08] *** djfirefox has quit IRC[18:45:27] *** WormDrink has quit IRC[18:46:18] *** sjorge has joined #openindiana[19:04:57] *** TomH has quit IRC[19:16:25] *** hunter has quit IRC[19:31:06] *** gwr has joined #openindiana[19:37:34] *** foton has joined #openindiana[19:37:50] *** troydm has joined #openindiana[19:44:48] *** foton has quit IRC[19:51:32] *** foton has joined #openindiana[19:52:40] *** randw has joined #openindiana[19:55:43] *** kovert1 has quit IRC[19:55:49] *** foton has quit IRC[19:55:52] *** kovert has joined #openindiana[20:05:29] *** nikolam has joined #openindiana[20:09:43] *** patdk-wk_ has joined #openindiana[20:11:26] *** patdk-wk has quit IRC[20:11:48] *** lblume has quit IRC[20:15:10] *** spanglywires has joined #openindiana[20:15:47] *** held has joined #openindiana[20:16:11] *** merzo has joined #openindiana[20:16:12] *** spanglywires has quit IRC[20:30:19] *** Savis has joined #openindiana[20:35:43] *** sjorge has quit IRC[20:45:45] *** randw_ has joined #openindiana[20:46:51] *** randw has quit IRC[20:49:29] *** movement has quit IRC[20:56:58] *** alex1 has quit IRC[20:57:02] *** randw_ has quit IRC[20:58:07] *** randw has joined #openindiana[20:58:24] *** Lee- has quit IRC[20:58:42] *** Lee- has joined #openindiana[20:59:54] *** Savis has quit IRC[21:04:17] *** hunter has joined #openindiana[21:06:24] *** drlou_on_tour has joined #openindiana[21:18:46] *** nikolam has quit IRC[21:30:10] *** ptribble has joined #openindiana[21:32:14] *** randw has quit IRC[21:37:25] *** irker408 has joined #openindiana[21:37:26] <irker408> spec-files-extra [5446] jurikm SFEpostfix.spec: sync usr/gnu/lib/sasl2 metadata with Cyrus SASL package[21:38:06] *** randw has joined #openindiana[21:39:42] *** heldus has joined #openindiana[21:41:47] *** merzo has quit IRC[21:41:54] *** held has quit IRC[21:49:05] <irker408> spec-files-extra [5447] jurikm SFEpostfix.spec: bump to 2.9.8[21:49:38] *** movement has joined #openindiana[21:51:03] *** movement has quit IRC[21:51:51] *** rfm has joined #openindiana[21:52:07] *** movement has joined #openindiana[21:56:35] <irker408> spec-files-extra [5448] tom68 TODO: add 73) request script to find conflicting files between repos early, add 74) re-visit svn diff -r 4965:5412 SFElibx264.spec to fix isaexec[22:03:06] *** nikolam has joined #openindiana[22:04:56] <irker408> spec-files-extra [5449] jurikm SFEmplayer2.spec: fix dependencies[22:05:15] *** gwr has quit IRC[22:05:21] *** WormDrink has joined #openindiana[22:10:36] <irker408> spec-files-extra [5450] jurikm SFElibx264.spec: read libtool cleanup[22:23:56] <irker408> spec-files-extra [5451] tom68 SFElivemedia.spec: rename to all lowecase package names (SFE standard)[22:29:04] *** drlou_on_tour has quit IRC[22:37:57] <irker408> spec-files-extra [5452] tom68 SFEcrrcsim.spec: change (Build)Requires to %{pnm_buildrequires_x11_library_freeglut}, %include packagenamemacros.inc[22:40:24] *** KermitTheFragger has quit IRC[22:40:56] <irker408> spec-files-extra [5453] tom68 SFEid3lib-gpp.spec: use std macros for *FLAGS, cc_is_gcc, remove SUNWlibC, add (Build)Requires: SFEgcc(runtime)[22:44:27] <irker408> spec-files-extra [5454] tom68 SFElibebml.spec: change BuildRequires to %{pnm_buildrequires_SUNWgnu_coreutils}[22:47:06] <irker408> spec-files-extra [5455] tom68 SFEvcdimager.spec: change (Build)Requires to %{pnm_buildrequires_SUNWlibpopt_devel} %{pnm_buildrequires_SUNWlxml_devel}, %include packagenamemacros.inc[22:50:46] <irker408> spec-files-extra [5456] tom68 ?\195?\130SFEr.spec: remove noisy comments, cleanup %files[22:50:53] *** movement has quit IRC[22:51:32] *** ptribble has quit IRC[22:52:46] *** gwr has joined #openindiana[22:56:26] *** rlaager has quit IRC[22:57:27] <irker408> spec-files-extra [5457] tom68 SFElibebml-gpp.spec: change BuildRequires to %{pnm_buildrequires_SUNWgnu_coreutils}, %include usr-g++.inc[23:04:45] <irker408> spec-files-extra [5458] jurikm encumbered/SFEmplayer2.spec: bump to the latest mplayer2 snapshot[23:07:37] <irker408> spec-files-extra [5459] tom68 include/osdistrofeatures.inc: temporary fix for IPS based OS for etc_security_directorylayout[23:08:27] <irker408> spec-files-extra [5460] tom68 SFEafterstep.spec: change (Build)Requires to %{pnm_buildrequires_library_readline}, %include packagenamemacros.inc[23:10:07] <irker408> spec-files-extra [5461] tom68 SFEffmpeg.spec: change BuildRequires to %{pnm_buildrequires_SUNWgsed} (S12)[23:13:56] <irker408> spec-files-extra [5462] tom68 encumbered/SFElame.spec: change (Build)Requires to %{pnm_buildrequires_SFElibsndfile_devel}, %{pnm_buildrequires_SUNWncurses_devel}, %include packagenamemacros.inc[23:15:56] <irker408> spec-files-extra [5463] tom68 SFEvice.spec: change (Build)Requires to %{pnm_buildrequires_library_readline}, %include packagenamemacros.inc[23:16:29] *** hunter has quit IRC[23:18:26] <irker408> spec-files-extra [5464] tom68 SFElibmpcdec.spec: change (Build)Requires to %{pnm_buildrequires_SFElibsndfile_devel}, %include packagenamemacros.inc[23:19:57] <irker408> spec-files-extra [5465] tom68 SFExmms2.spec: change (Build)Requires to %{pnm_buildrequires_SUNWlibdiscid_devel}, %include packagenamemacros.inc[23:20:24] *** sjorge has joined #openindiana[23:21:37] <irker408> spec-files-extra [5466] tom68 SFEtwolame.spec: change (Build)Requires to %{pnm_buildrequires_SFElibsndfile_devel}, %include packagenamemacros.inc[23:21:46] <irker408> spec-files-extra [5467] viskov SFEmplayer2.spec: make %snap date of latest commit to the git repository, to ease maintenance[23:23:56] <irker408> spec-files-extra [5468] tom68 SFExmms1-scrobbler.spec: change (Build)Requires to %{pnm_buildrequires_SUNWlibmusicbrainz_devel}, %include packagenamemacros.inc[23:25:07] <irker408> spec-files-extra [5469] tom68 SFEid3lib-gnu.spec: align Summary with SFEid3lib-gpp.spec add URL[23:25:56] <irker408> spec-files-extra [5470] tom68 SFElibtunepimp.spec: change (Build)Requires to %{pnm_buildrequires_library_readline}[23:31:15] <irker408> spec-files-extra [5471] tom68 SFEtumbler.spec: bump to 0.1.29, use ffmpeg anytime, change to BuildRequires: %{pnm_buildrequires_SUNWgtk_doc}, %include packagenamemacros.inc[23:31:17] *** blues_air has joined #openindiana[23:33:00] *** blues_air has quit IRC[23:34:47] <irker408> spec-files-extra [5472] tom68 SFEnagios.spec: change (Build)Requires to %{pnm_buildrequires_perl_default}, export CC=gcc, CXX=g++, CFLAGS with perl, cc_is_gcc = 1[23:42:06] <irker408> spec-files-extra [5473] tom68 SFEreadline.spec: move to /usr/gnu to avoid duplicate files with OS provided readline, add IPS_Package_Name, fix %files[23:46:53] <tsoome> spam:P[23:48:22] *** hunter has joined #openindiana[23:54:26] <irker408> spec-files-extra [5474] tom68 SFEnaim.spec: change (Build)Requires to %{pnm_buildrequires_SUNWncurses_devel}, %include packagenamemacros.inc[23:58:27] <irker408> spec-files-extra [5475] tom68 SFEghc.spec: change (Build)Requires to %{pnm_buildrequires_SUNWncurses_devel}