July 20, 2011 
< | 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 | 31 | >
July 20, 2011 [00:12:08] *** DrLou has quit IRC[00:22:47] *** merzo has joined #openindiana[00:34:38] *** thedeo has quit IRC[00:36:22] *** InTheWings has quit IRC[00:39:40] *** merzo has quit IRC[00:41:58] *** gber has quit IRC[01:03:32] *** DrLou has joined #openindiana[01:03:32] *** ChanServ sets mode: +o DrLou[01:09:16] *** Alasdairrr is now known as AlasAway[01:16:04] *** gea has quit IRC[01:24:09] *** DontKnwMuch has joined #openindiana[01:25:35] <DontKnwMuch> How can I find a cause of cpu usage growing 3% per day, I was at 30% and after reboot, everything was still working but cpu was 2%[01:26:13] *** axisys_ has quit IRC[01:26:28] *** axisys has joined #openindiana[01:42:17] *** kdavy_ has joined #openindiana[01:44:06] *** kdavy__ has quit IRC[01:50:23] *** kdavy_ has quit IRC[02:03:09] *** CME has quit IRC[02:07:05] *** CME has joined #openindiana[02:43:45] *** winIO has quit IRC[02:48:58] *** arpunk has joined #openindiana[02:48:59] *** redgone has joined #openindiana[02:58:58] *** jasonmiu has quit IRC[03:03:09] *** xmikus01 has quit IRC[03:11:56] *** master_of_master has quit IRC[03:13:50] *** master_of_master has joined #openindiana[03:17:12] *** held has quit IRC[03:24:25] *** held has joined #openindiana[03:30:09] *** skeeziks has joined #openindiana[03:37:19] *** DrLou has quit IRC[03:47:46] *** LizardKing has quit IRC[03:56:45] *** swa_work has joined #openindiana[04:10:48] *** magyar has joined #openindiana[04:10:48] *** magyar has joined #openindiana[04:12:30] *** CME has quit IRC[04:12:31] *** LizardKing has joined #openindiana[04:26:07] *** redgone has quit IRC[04:49:14] *** POloser has joined #openindiana[04:50:21] *** magyar has quit IRC[05:05:05] *** CME has joined #openindiana[05:08:48] *** Naresh has quit IRC[05:10:29] *** McBofh has quit IRC[05:14:01] *** kart_ has joined #openindiana[05:18:02] *** Transformer has joined #openindiana[05:27:28] *** kdavy_ has joined #openindiana[05:29:12] *** smrt has quit IRC[05:29:31] *** smrt has joined #openindiana[05:45:29] <kdavy_> DontKnwMuch, have you identified the process that is causing this? You can do that using "prstat"[05:45:39] <POloser> how passed developer meeting? any news on oi_151 release date?[05:46:03] * kdavy_ is not a developer...[05:47:57] *** Norrin has quit IRC[05:48:15] *** Norrin has joined #openindiana[06:18:53] *** DeanoC has quit IRC[06:19:39] *** DeanoC has joined #openindiana[06:19:39] *** ChanServ sets mode: +o DeanoC[06:40:31] *** baitisj has quit IRC[06:54:35] *** swa_work has quit IRC[07:01:18] *** baitisj has joined #openindiana[07:02:34] <baitisj> Hey guys, question. I want some of my interfaces to *NOT* have the router flag on reboot.[07:02:50] <baitisj> Can I change e.g. /etc/hostname.vnic0 to say:[07:02:55] <baitisj> 192.168.2.3 -router[07:03:18] <richlowe> I think that should work[07:03:37] <richlowe> but I'm almost always wrong about how forgiving hostname.if files are[07:04:02] <richlowe> if I remember correctly, you should ask me what I think the me that's always right would say, and then do the opposite ;)[07:09:02] <madwizard> Coffee[07:11:32] *** LeftWing has quit IRC[07:14:52] *** LeftWing has joined #openindiana[07:21:27] *** Naresh has joined #openindiana[07:26:16] *** tsukasa has quit IRC[07:28:59] <blues> anyone doing multipath iscsi between their OI box and esxi host?[07:32:07] *** McBofh has joined #openindiana[07:33:43] *** kdavy_ has quit IRC[07:39:10] <c00p> anyone got their oi box using radius for the shadow / password and /etc/passwd for account info ?[07:39:37] <c00p> got a radius system now with a otp system (MiToken) that I want to point my oi box at[07:39:45] *** tsukasa has joined #openindiana[07:45:03] *** skeeziks has quit IRC[07:48:27] *** n2deep has quit IRC[07:48:46] *** n2deep has joined #openindiana[07:58:39] *** mnaser has quit IRC[08:12:34] *** DontKnwMuch has quit IRC[08:16:03] *** flyz has joined #openindiana[08:22:58] *** Abbaduxo has joined #openindiana[08:27:49] <sivanov> hmm[08:27:56] <sivanov> Children see throughput for 4 readers = 394830.70 KB/sec[08:28:07] <sivanov> Children see throughput for 4 re-readers = 82228.61 KB/sec[08:28:18] <sivanov> Children see throughput for 4 random readers = 269732.80 KB/sec[08:29:01] <sivanov> reading from hot cache: 390MB/s then re-read @82MB/s the ramdom read @269MB/s. over 4GB FC[08:32:29] *** kart_ has quit IRC[08:33:39] *** kart_ has joined #openindiana[08:33:53] *** McBofh has quit IRC[08:38:40] *** McBofh has joined #openindiana[08:38:59] *** gea has joined #openindiana[08:44:50] *** baitisj has quit IRC[08:48:03] *** LeftWing has quit IRC[08:48:03] *** syoyo_ has joined #openindiana[08:48:17] *** DeanoC has quit IRC[08:48:31] *** |AbsyntH| has joined #openindiana[08:48:46] *** LeftWing has joined #openindiana[08:49:08] *** eryc has quit IRC[08:49:22] *** eryc has joined #openindiana[08:49:52] *** lkocman has joined #openindiana[08:50:06] *** Botanic has quit IRC[08:50:22] *** kovert has quit IRC[08:51:03] *** Botanic has joined #openindiana[08:51:47] *** kovert has joined #openindiana[08:53:50] *** DeanoC has joined #openindiana[08:53:50] *** ChanServ sets mode: +o DeanoC[08:56:20] *** Abbaduxo has quit IRC[08:56:20] *** merzo has joined #openindiana[08:57:12] *** Worsoe has joined #openindiana[08:59:46] *** DontKnwMuch has joined #openindiana[09:02:37] <DontKnwMuch> hi :) I have an interesting problem, my CPU usage in top is increasing gradually approx 3% per day of uptime. When it reached 30 or so, OI hung. After a reboot it is very low, but now after 34 hours it is 7% already. How can I find a cause?[09:03:17] <Triskelios> prstat, intrstat[09:06:07] <DontKnwMuch> in intrstat everything is actually almost 0% for each cpu, prstat the same, the highest cpu usage os 0.2% for zpool[09:09:29] <DontKnwMuch> os = is[09:12:03] *** easyE has joined #openindiana[09:12:07] *** bens1 has joined #openindiana[09:16:06] *** baitisj has joined #openindiana[09:27:02] *** tsoome has quit IRC[09:30:18] *** Botanic has quit IRC[09:36:21] *** Micr0mega has joined #openindiana[09:46:05] *** TachyonC has joined #openindiana[09:46:21] *** TachyonC has quit IRC[09:48:37] *** Botanic has joined #openindiana[09:53:08] <DontKnwMuch> prstat -s cpu -n 5 shows: http://www.pastie.org/2241576[09:53:47] <DontKnwMuch> mpstat does have some cpu usage: http://www.pastie.org/2241586[09:54:32] *** kart_ has quit IRC[09:57:03] *** tsoome has joined #openindiana[10:00:22] *** lkocman has quit IRC[10:04:35] <DontKnwMuch> what can be in kernel that is gradually consumming more and more cpu, and how can i find it? prstat does not show any usage.. as usage is in kernel..[10:04:59] <tsoome> sys time is increasing?[10:05:42] <DontKnwMuch> in mpstat yes[10:05:54] <quasi> prstat -m[10:05:56] <DontKnwMuch> cpu 0 sys is higher and higer[10:05:59] <tsoome> pastebin some amount of samples from mpstat 1[10:06:13] <DontKnwMuch> prstat -s cpu -n 5 shows: http://www.pastie.org/2241576[10:06:18] <DontKnwMuch> mpstat does have some cpu usage: http://www.pastie.org/2241586[10:06:27] *** kart_ has joined #openindiana[10:07:31] *** syoyo_ has quit IRC[10:08:37] <tsoome> run /opt/DTT/hotkernel for some time[10:10:21] *** Phenox_ has quit IRC[10:11:36] *** Micr0mega has left #openindiana[10:11:46] *** Micr0mega has joined #openindiana[10:18:02] <DontKnwMuch> tsoome: it hanged it... it still responds to ping... but I can not access it...[10:18:20] <tsoome> .oO[10:18:24] <DontKnwMuch> "high TCP connect timeout rate! System[10:18:36] <DontKnwMuch> (port 139 may be under a SYN flood attack![10:18:39] <DontKnwMuch> :([10:18:41] <DontKnwMuch> argh...[10:19:05] *** c00p has quit IRC[10:19:06] <tsoome> you may need firewall;)[10:19:21] <tsoome> what is reporting that message?[10:19:35] <DontKnwMuch> its on the console[10:19:48] <tsoome> nice[10:20:05] <DontKnwMuch> it hung completely[10:20:16] <tomww> this is kernel CIFS, I've never seen such with samba[10:20:20] <tomww> ?[10:21:04] <DontKnwMuch> let me reboot it... I would kill the hotkernel which I run just before it hung[10:22:28] <DontKnwMuch> I had to reset it by ikvm...[10:22:43] <tsoome> didnt you had some issues with cifs before as well?[10:22:52] <DontKnwMuch> no[10:23:04] <tsoome> ah, must have been someone else[10:23:06] <DontKnwMuch> I only had higher and higer CPU usage[10:23:16] <DontKnwMuch> 3% per day increase[10:23:49] <DontKnwMuch> I also had one hang a few days ago when it reached 30%[10:23:56] <DontKnwMuch> in top kernel consupmtion that is[10:24:01] <tsoome> well, still i suggest to get some samples with hotkernel from time to time, it should reveal what is taking the time[10:24:31] <DontKnwMuch> I am sortof afraid to run it again ;) I sould not hang my system otherwise[10:24:55] <tsoome> in normal situation it wont do that[10:25:18] <tsoome> as it did, its already an sign something is very wrong in your system[10:26:34] <tsoome> hotkernel is basically just counting how many times the functions have been called during its run time[10:29:17] <tomww> does a broken network card of a cifs client constantly re-try connecting?[10:29:38] <tsoome> you can see that from snoop[10:31:34] <DontKnwMuch> ok, I had to answer some 30 phone calls ;([10:31:39] <DontKnwMuch> its back up[10:31:46] <DontKnwMuch> but damn is this stressfull[10:31:48] <tsoome> also check netstat as well[10:32:56] <tsoome> um, someone here had high system load caused by one single misbehaving cifs client[10:33:28] <tsoome> but he didnt had hungs, just high load[10:33:37] <DontKnwMuch> could be...[10:34:30] <tsoome> so it may be good idea to check network traffic as well, if there is some kind of abnormal pattern[10:35:20] <tsoome> as well as traffic from systems which shouldnt be there (if that syn flood message was correct;)[10:35:44] <DontKnwMuch> I did run dladm show-link -s and had 0 errors on both network cards in the system just before[10:35:52] <DontKnwMuch> it had some 30 hours uptime[10:36:19] <DontKnwMuch> so I think network cards are not broken[10:37:12] <tsoome> that is for physical errors the phys layer can detect[10:38:22] <tsoome> also, the server connection can be just fine, but misbehaving client can still cause trouble[10:38:25] <DontKnwMuch> this syn flood high TCP connect timeout could be because it really was timing out, as CPU consumption was 100% or something[10:38:25] <tomww> which program was this printing the current tcp connections with traffix, so the equivalent of linux's iptraf ...[10:39:02] <tsoome> there are loads of scripts in DTT[10:39:26] <DontKnwMuch> this one I am afraid to run again ;)[10:40:01] <tomww> btw, the DTT wasn't updated for a long time, does anyone collect updated stripts so we could update the regular old DTT with them ourselfes?[10:40:02] <tsoome> well, if running dtrace will break your system, it only means it was dead already[10:40:36] <DontKnwMuch> after reboot my kernel in top is 0.4%, it was 9% before[10:41:11] <tsoome> its an shame apple did include the DTT scrips in base os, but in solaris they are still like stepchild:([10:42:15] *** POloser has left #openindiana[10:42:29] <tsoome> Nazgul:~ tsoome$ man -k dtrace|wc -l[10:42:30] <tsoome> 47[10:45:00] <DontKnwMuch> an hour before I had this in log: http://www.pastie.org/2241787 what could that be?[10:46:36] *** man_u has joined #openindiana[10:48:34] *** POloser has joined #openindiana[10:52:47] *** Botanic has quit IRC[10:55:30] <DontKnwMuch> tomww: Can I actually find/see which client is constantly retrying connecting? Could be as you said, a broken network card of a cifs client.[10:55:58] <DontKnwMuch> snoop? ah...[10:57:36] <lblume> or netstat -an and look at a bunch of connections from the same IP[11:00:41] *** gea has quit IRC[11:06:11] <RoyK> "Typical asshole Porsche driver who takes up two parking spots!": http://www.youtube.com/watch?v=K-abIbjdZZM[11:08:34] <DontKnwMuch> wow, he parked it fast :)[11:09:47] *** xmikus01 has joined #openindiana[11:31:37] *** held has quit IRC[11:32:52] *** andygraybeal has quit IRC[11:41:22] <tomww> DontKnwMuch: about our http://www.pastie.org/2241787 , this look slike just that there have been loaded a bunch of modules, and this prints the pseudo devices the modules used. level is "info", so nothing to worry about.[11:42:33] <tsoome> this is normal output from kernel[11:51:27] *** AlasAway is now known as Alasdairrr[11:53:17] *** mikw has joined #openindiana[11:53:51] *** anikin has joined #openindiana[11:54:04] *** mikw has quit IRC[11:54:37] *** mikw has joined #openindiana[11:56:35] <DontKnwMuch> tomww: thanks, I just went for a walk to cool down my stress levels ... ;)[11:58:17] <DontKnwMuch> tomww: but why is it happening is a mistery for me, I did not have it in my logs until today, and now it repeats itself almost every half hour[12:00:10] <DontKnwMuch> interesting is that I have a clean install, nothing added, just kernel smb and nfs shares, so probably the problems I am having are caused by external factors...[12:03:06] <tomww> if you have a simple find /dev, or a devfsadm scanning al possible devices, then the kernel loads those drivers who's devicenodes are opened.[12:04:01] <tomww> can be a backup software or any other program. even firefox triggers this by looking ad /dev entries (I believe)[12:04:13] <tsoome> dtrace most definitely will[12:05:11] <DontKnwMuch> oh.. ok[12:07:35] <tomww> I cleanup from time to time (this is after I remove my CF-card or USB-memorystick the /dev entries with devfsadm -Cv[12:08:25] <tomww> avoid long wait times on dev entries from physicly removed storage media[12:08:58] <DontKnwMuch> Will write this one down[12:09:27] *** held has joined #openindiana[12:10:54] <DontKnwMuch> I will try to run the /opt/DTT/hotkernel in the evening when no one is in the office, to see what will happen[12:11:14] <DontKnwMuch> and perhaps also see what is consuming my cpu[12:11:56] <DontKnwMuch> it now almost two hours after reboot and it went from really no load to some 1.5-2% for kernel[12:12:15] <DontKnwMuch> and its steadily growing[12:13:02] *** pjama has left #openindiana[12:13:08] *** mnaser has joined #openindiana[12:17:24] <|woody|> tomww for new Dtrace scripts to use just check http://www.dtracebook.com/index.php/Main_Page[12:18:56] <DontKnwMuch> these http://www.dtracebook.com/index.php/System_View simple one liners are nice :)[12:20:08] *** |AbsyntH| has quit IRC[12:22:59] <DontKnwMuch> I will try a different approach. I will just stop poking at it, and leave it be. If it does not hang in a few days, I am ok ;)[12:23:27] <DontKnwMuch> woody: this is a nice book[12:24:08] <|woody|> of couse it is[12:26:04] <DontKnwMuch> a very nice one indeed :)[12:26:06] *** mnaser has quit IRC[12:27:25] <|woody|> I actually should start to read it. But didn't have the time yet[12:41:07] *** lennard has quit IRC[12:46:06] *** lennard has joined #openindiana[12:46:51] <taemun> http://pastie.org/private/ci9fe1qimvauzjydxeieiq[12:46:57] <taemun> have a look at the third last second there[12:47:01] <taemun> any ideas what happened?[12:51:30] <tsoome> bug maybe from counter switch[12:52:10] *** InTheWings has joined #openindiana[13:00:15] <DontKnwMuch> hm... this is a similar thing... well sortof... http://web.archiveorange.com/archive/v/9QxCTtfcPslXEyo3RkQR I will disable power management to see what happens[13:08:51] <DontKnwMuch> I will also disable all the not needed crap in bios, lik VT-x etc..[13:10:19] *** syoyo_ has joined #openindiana[13:19:03] *** heldchen has joined #openindiana[13:20:26] *** heldchen has quit IRC[13:21:11] *** held has quit IRC[13:23:26] *** sudosu has joined #openindiana[13:40:47] *** mikw has quit IRC[13:50:12] *** tsoome1 has joined #openindiana[13:51:02] *** tsoome has quit IRC[13:51:02] *** tsoome1 is now known as tsoome[13:55:26] *** tsoome has quit IRC[13:59:15] *** tsoome has joined #openindiana[14:02:49] <tsoome> .oO oh, we have that very important system broken, it must be repaired asap. no, we wont pay for support, its too expensive. damn, i "like" such clients...[14:02:58] <viridari> Good morning, all. The Google is not my friend today. Trying to figure out, with oi as a guest under ESXi 4.1, how to install oi in such a way that it doesn't try to use a serial port console.[14:03:46] <viridari> it's possible to use a serial port console for vmware guests, but a pain in the butt for managing a large number of guest OS's. Would rather use the VMware console like we do for all other guests.[14:07:55] <DontKnwMuch> viridari: I have it installed, but can use vmware console normally, what (how) are you doing it?[14:08:34] *** POloser has left #openindiana[14:09:43] <DontKnwMuch> tsoome: was this a point to me?[14:09:58] <viridari> DontKnwMuch: I installed from live CD, the VMware console worked fine there. Upon rebooting after the initial installation, I get the GRUB loading stage 2 message and then nothing further on the VMware console. I might be making an assumption here that the serial console took over, but that seems consistent with what I've seen on Solaris hosts that really are using serial console.[14:11:58] <DontKnwMuch> viridari: this is strange, it should just work as it is. Do you have some vmdirectpath or something..,[14:12:47] <viridari> I don't even know what a vmdirectpath is. I told ESXi that I was installing Solaris 10 64-bit (closest template I could find), added a second VCPU, bumped up memory and disk. Otherwise took defaults.[14:12:49] <DontKnwMuch> are you accessing it locally? the vmware console that is[14:13:11] <DontKnwMuch> or are you remote (on wan)[14:13:21] <viridari> remote[14:13:22] <DontKnwMuch> with locally I mean lan[14:13:43] <viridari> well, the ESX server is about 30 feet from me :) but on a different subnet[14:13:47] <DontKnwMuch> so you are over ipsec or smoething[14:13:48] <DontKnwMuch> ah...[14:14:00] <DontKnwMuch> this is a problem than[14:14:14] <viridari> oh? all other OS's seem to work fine this way[14:14:18] <viridari> as guests I mean[14:14:19] <DontKnwMuch> the console has some strange requirements... well... hm... if other consoles work...[14:15:01] <viridari> oh yeah I've got other VMware servers in another county, managing them via forwarded ports over ssh, consoles work fine on them, too... with Windows, Linux, and FreeBSD guests[14:15:03] <DontKnwMuch> I had a problem with everything but consoles work, as some ports or something were blocked, but if other consoles work, I am clueless[14:15:17] <viridari> and it worked fine for the installation itself[14:17:04] <viridari> in a perfect world I could install from the oi Server CD and not have it default to a serial console... no need for the GUI bits in my VM's. But I'm instead installing from the Live disc just to avoid the serial console... and failing. I'm guessing there is some hidden bit that I'm missing.[14:17:06] <DontKnwMuch> are there any errors in vmware[14:17:15] <viridari> no[14:17:40] <viridari> that VM was nuked & paved, I installed FreeBSD in its place just to get the MySQL server up that I needed. Going to spin up another one now.[14:17:45] <DontKnwMuch> just install from oi server cd, I have installed 3 this way, and they just worked[14:17:57] <DontKnwMuch> no GUI[14:18:04] <viridari> ok will try that now[14:20:53] *** c00p has joined #openindiana[14:21:09] <viridari> grrr I deleted the server disc image... might be awhile (re-downloading)[14:23:18] <viridari> lol ok that's going to be hard to do via bittorrent with zero seeders ;)[14:23:30] <viridari> oops spoke too soon[14:27:45] <gaYak> viridari: If you get stuck on "loading stage 2", then it's not serial port.. it's dead VM.[14:27:49] <tsoome> DontKnwMuch: no, one of our clients;)[14:28:37] <DontKnwMuch> oh thank god ;)[14:28:44] <gaYak> viridari: At least my OI died everytime with ESXi, if I used physical disk mapping[14:28:51] <DontKnwMuch> I would put "paying customers only" on the door, lol[14:29:40] <viridari> gaYak: interesting... the machines where I have Solaris 11 and OI running on iron, that's the last message I would see on an attached VGA monitor before everything is only visible via the management port[14:29:52] <viridari> (hence my assumption about what was happening)[14:30:53] <gaYak> viridari: Right, well when I tried to install OI to our ESXi, I stumbled to that GRUB stage problem and the system never finished from there. After mm.. many reinstalls and partitioning tricks and reinstalling GRUB and what not, I found out the pdm was corrupting the disks[14:31:02] <gaYak> (our ESXi machine does not support VT-d)[14:31:56] <DontKnwMuch> VT-d does work very good, pdm has problems, all over the place, not only on oi[14:32:13] <gaYak> DontKnwMuch: Yeah, I wasn't saying it was OI's problem but rather ESXi's[14:32:44] <gaYak> DontKnwMuch: Anyway, our ESXi host didn't unfortunately have that feature. So I eventually used virtual disk mapping (I didn't actually even notice any performance difference)[14:32:47] <gaYak> And it works fine[14:33:12] <viridari> interesting[14:33:17] <DontKnwMuch> The only problem on esxi is that if more then one vcpu is added to oi, the esxi cpu consumption goes to 100% on any higher then idle load[14:33:36] <gaYak> huh?[14:33:40] <viridari> d'oh (looking at fresh VM with 2x VCPU)[14:34:08] <DontKnwMuch> a OI vm? put it on some load, copy 5gb or so and wathc cpu usage in esxi[14:34:14] <gaYak> I have 2 or 4 vCPU.. let me check[14:34:14] *** taemun has quit IRC[14:34:20] <gaYak> 4 vCPUs (and 8GB RAM)[14:35:01] <gaYak> Sure, moment[14:35:14] *** taemun has joined #openindiana[14:36:11] <tsoome> .oO OSX Lion Downloading - 10 minutes remaining....[14:36:23] <viridari> tsoome: \o/[14:36:31] <viridari> tsoome: looking forward to hearing your initial impressions[14:36:37] <gaYak> My streaming test seems to get high it (eats about ~6GHz), but not to quite 100%[14:36:52] <tsoome> 3.74GB[14:36:52] <DontKnwMuch> check in top how much cpu is used there[14:37:22] <DontKnwMuch> whenever I tried it was 20 or so, bit in esxi it was almost quite 100%[14:37:33] <DontKnwMuch> when changed to one vcpu it is normal as expected[14:37:47] <DontKnwMuch> bit = but[14:38:08] <DontKnwMuch> used there = in oi[14:38:13] <gaYak> I can't get it to higher than 70% kernel[14:38:37] <gaYak> But I admit, that's way more than it should eat[14:38:38] <DontKnwMuch> and in esxi for this vm, cpu usage graph[14:38:48] <gaYak> Shows about the same[14:39:01] <gaYak> Well ok it topped at 92%[14:39:06] <DontKnwMuch> way too much[14:39:13] <DontKnwMuch> change to one vcpu and test again[14:39:20] <DontKnwMuch> when you will be able to shut it down[14:39:32] <gaYak> Now that's good question.. that takes down all the VMs.. ;)[14:39:58] <DontKnwMuch> it is a bug somewhere in esxi/oi or something, w2k8 guests running on more vcpus use normal amounts of cpu on esxi[14:40:22] <gaYak> Yea, I mostly have W2K8R2 hosts.. though I planned to move them to our Hyper-V host[14:40:25] <gaYak> .. guests.[14:40:33] <viridari> yeah I have FreeBSD guests mostly and with multiple VCPU, utilization seems normal[14:40:49] <gaYak> But that's irritating bug.. I'm running more than just ZFS sharing on my OI.[14:40:51] <DontKnwMuch> oi and more vcpus just dont work good, it even works faster with one vcpu[14:41:27] <DontKnwMuch> iritating is a mild word[14:42:24] <DontKnwMuch> I went back to one vcpu and since its a nfs/smb and nothing more, it runs way faster and uses less cpu on esxi too[14:45:23] <viridari> ok yeah my brand new oi_148 server installation under ESXi 4.1 boots as far as the copyright notice and then stops[14:45:49] <DontKnwMuch> how long.. it takes some time usually....[14:46:06] <viridari> will let it sit awhile[14:48:10] *** DrLou has joined #openindiana[14:48:10] *** ChanServ sets mode: +o DrLou[14:48:41] *** setupji_ has quit IRC[14:50:25] *** tsoome has quit IRC[14:51:22] <DontKnwMuch> well, viridari... this is long enough I beleive[14:51:46] <viridari> still stuck at copyright banner[14:52:18] <viridari> three lines on the screen: OpenIndiana build number, SunOS Release banner, All rights reserved banner[14:54:00] <tomww> add "-v" to the grub line with kernel on it[14:54:17] <tomww> then you can see which modules have been loaded successfully[14:54:43] * viridari reboots with tomww's advice[14:54:51] *** tsoome has joined #openindiana[14:55:01] *** tsoome has quit IRC[14:55:11] *** tsoome has joined #openindiana[14:56:06] <viridari> well it is loading smf service descriptions so far and still going[14:56:19] *** held has joined #openindiana[14:56:40] <viridari> o_O I have a login prompt[14:57:15] *** tsoome_ has joined #openindiana[14:57:15] *** tsoome has quit IRC[14:57:15] *** tsoome_ is now known as tsoome[14:58:26] *** ivwang has joined #openindiana[14:58:53] <tsoome> Hm, bloody 30min install...[14:59:43] <viridari> ok this is odd[15:00:25] <viridari> if I power down the VM and bring it back up without mucking in GRUB, it comes up. So it's only getting stuck if I reboot after the install without powering the VM off first.[15:00:27] *** hsp has joined #openindiana[15:01:37] <gaYak> dontknwmuch: Lets see.. everyone left work, so I was able to shut it down (lucky summer)[15:02:40][15:03:35] <gaYak> What mm.. it got different IP.[15:04:07] <DontKnwMuch> ugh ugh ... I just hate iscsi[15:04:36] <gaYak> Well, NFS doesn't work that well in ESXi either.[15:04:43] <gaYak> Performance is quite .. horrible.[15:05:03] <DontKnwMuch> NFS with ssd log /cache is quite ok[15:05:27] <DontKnwMuch> well.. I did not benchmark it...[15:05:36] *** smrt has quit IRC[15:05:43] <gaYak> I actually do have SSD log & cache[15:05:46] <gaYak> But still wasn't happy[15:05:54] *** smrt has joined #openindiana[15:06:11] <DontKnwMuch> oh.. iscsi works how much faster for you?[15:06:11] <gaYak> It was much slower than iSCSI<->ESXi[15:06:19] <gaYak> About twice as fast[15:06:37] <tsoome> Make sure the tcp tunings are done[15:06:46] <viridari> ok with 2x VCPU, I'm hovering at 99.7% idle[15:07:30] <DontKnwMuch> viridari: try to put some load on it[15:07:48] <DontKnwMuch> tsoome: you are using the tunings form that page for you setups?[15:08:28] <tsoome> In some, not all. Need to test them more[15:08:50] <DontKnwMuch> I was not courageous enough to try any ;)[15:09:23] <tsoome> Cifs perfomance diff in lan was 2x[15:09:49] <DontKnwMuch> nfs?[15:10:04] <tsoome> Havent tested[15:10:38] <DontKnwMuch> will try some , sounds good[15:10:40] <gaYak> Maybe NFS would've required a better SSD, but at least Intel's X25-M wasn't enough speedy.[15:10:48] <tsoome> Also the large windows mean it will help large transfers[15:11:14] <gaYak> Then again.. if iSCSI works, why switch.[15:11:59] <DontKnwMuch> Funny thing is that when I tired the machine before deploying, from one client I was getting full glan speed on cifs, now, when put on the network, it is 50-70mb/s max, even when there is no load[15:12:19] <tsoome> If you got it still on nfs, zfs set sync=disabled and see what happens:P[15:13:33] <tsoome> Another thing is, you wanna make many mounts, not just one huge[15:13:52] <gaYak> Reason being..?[15:13:54] <viridari> DontKnwMuch: is that reads or writes?[15:14:00] <tsoome> But thats same with iscsi as well[15:14:02] <DontKnwMuch> both more or less[15:14:10] *** |AbsyntH| has joined #openindiana[15:14:19] <viridari> DontKnwMuch: using zfs based iscsi targets, I saw some big differences between read & write performance[15:14:34] <tsoome> Multiple connections[15:14:55] <DontKnwMuch> I have 4 mounts or so, and I aggregated two nics, but the load is distributed 80:20 on them, nfs not iscsi[15:15:23] <tsoome> Mounts from different datasets?[15:15:24] <gaYak> tsoome: With multiple nics sure.. but with one virtual 10Gbps, network performance shouldn't be case?[15:15:33] <DontKnwMuch> dladm show-aggr -s that is, yes different[15:15:37] <tsoome> It is actually[15:16:16] <tsoome> Because with single connection you are serializing io calls[15:16:28] <gaYak> DontKnwMuch: 1 vCPU = 100% CPU usage...[15:17:44] <DontKnwMuch> and speed is the same?[15:19:02] <gaYak> Well, read performance was.. I forgot to save my write results.. ;)[15:19:38] <DontKnwMuch> oh ...[15:21:03] *** anikin has quit IRC[15:21:10] <DontKnwMuch> and how much cpu usage in oi itself, the same?[15:21:24] *** anikin has joined #openindiana[15:24:46] <DontKnwMuch> the interesting thing is that it was using much more cpu (hz) for the same thing when more than one vcpu was used.. it even used twice more.. at least for me[15:36:07] <gaYak> DontKnwMuch: In OI, it showed higher kernel numbers..[15:40:23] *** Phenox_ has joined #openindiana[15:44:35] *** skeeziks has joined #openindiana[15:48:01] *** mnaser has joined #openindiana[15:49:10] *** kdavy_ has joined #openindiana[15:49:13] <viridari> there was some chatter here this morning about OS X Lion. If you're upgrading to Lion, you will find this useful: http://www.macrumors.com/2011/07/18/make-an-os-x-lion-boot-disc/[15:49:31] *** Micr0mega has left #openindiana[15:50:02] *** sudosu has quit IRC[15:51:04] *** tsoome1 has joined #openindiana[15:53:37] *** Worsoe has quit IRC[16:03:10] *** sudosu has joined #openindiana[16:10:57] *** syoyo_ has quit IRC[16:25:12] *** tsoome1 has quit IRC[16:33:48] *** Naresh has quit IRC[16:34:43] <Infin1ty|work> trying to build a manifest for AI, but it complains about "destination" in the software tags, i could not find anything by looking online, i just came across this: http://www.mail-archive.com/install-discuss at opensolaris dot org/msg00501.html, the "destination" is not documented anywhere, that includes oracle documents as well[16:35:52] *** tsoome has quit IRC[16:38:21] *** gea has joined #openindiana[16:41:19] * viridari covets http://www.newegg.com/Product/Product.aspx?Item=N82E16820227659 for his home OI box[16:42:17] <Infin1ty|work> Element software content does not follow the DTD, expecting (destination? , source* , software_data*), got (source name software_data software_data ) , "destination"?[16:48:06] *** anikin has quit IRC[16:55:48] *** raichoo has joined #openindiana[16:56:03] <raichoo> hi folks[16:57:18] *** |AbsyntH| has quit IRC[16:58:08] *** tsukasa has quit IRC[16:58:10] *** |AbsyntH| has joined #openindiana[16:58:50] *** ivwang has quit IRC[17:10:56] *** tsoome has joined #openindiana[17:10:58] *** tsoome_ has joined #openindiana[17:12:10] *** tsukasa has joined #openindiana[17:14:11] *** sudosu has quit IRC[17:19:46] *** sudosu has joined #openindiana[17:28:04] *** miine has quit IRC[17:33:44] *** merzo has quit IRC[17:34:46] *** miine has joined #openindiana[17:35:27] *** bens1 has quit IRC[17:49:14] <mnaser> does solaris/openindiana offer a remote api by any chance? (specifically for zfs and/or stmfadm?)[17:49:14] <mnaser> o[17:49:19] <mnaser> or will i need to ssh in, etc[17:51:05] *** auronandace has quit IRC[17:51:31] *** auronandace has joined #openindiana[17:54:09] *** kytibe has quit IRC[17:57:08] *** kytibe has joined #openindiana[17:59:41] <tsoome> no api[17:59:47] <mnaser> aww[18:00:16] <tsoome> but there are some web based management tools[18:00:30] <mnaser> i already have an implementation that sshs in, etc[18:21:47] *** syoyo_ has joined #openindiana[18:30:48] *** setupji has joined #openindiana[18:31:51] *** forquare has joined #openindiana[18:40:00] *** |AbsyntH| has quit IRC[18:40:25] *** forquare has quit IRC[18:53:25] *** Botanic has joined #openindiana[19:00:16] *** doug_ndndn has quit IRC[19:23:11] *** doug_ndndn has joined #openindiana[19:25:26] *** syoyo_ has quit IRC[19:30:35] *** Naresh has joined #openindiana[19:42:47] *** Botanic has quit IRC[19:46:16] *** baitisj has quit IRC[20:02:45] *** held has quit IRC[20:10:50] *** merzo has joined #openindiana[20:22:59] *** arpunk has quit IRC[20:27:59] *** Hedonista_ is now known as Hedonista[20:28:07] *** tsukasa has quit IRC[20:28:22] <gea> Does anyone have also tried to install vmware-tools (ESXi 4.1) on a german version of OI 151? It fails with a lot of perl errors.[20:28:24] <gea> Installation on a english version of OI 151 is ok.[20:29:12] * RoyK has never tried a non-english version of OI[20:29:31] * dkeav either[20:29:42] <dkeav> i'm not very good with german[20:30:53] <gea> maybee a problem with all non-english versions - i have tried two machines several times always the same result[20:32:55] *** jedi4ever has joined #openindiana[20:37:23] *** held has joined #openindiana[20:39:51] *** Botanic has joined #openindiana[20:39:51] *** Botanic has joined #openindiana[20:40:12] *** tsukasa has joined #openindiana[20:44:54] *** jpg has quit IRC[20:55:07] <tsoome> you mean, non english installer?[20:55:49] <tsoome> isn't that issue with locales using , instead of . in numbers?[20:56:56] *** hsp has quit IRC[20:57:45] *** hsp has joined #openindiana[20:59:09] *** kart_ has quit IRC[20:59:44] <gea> Installing with default installer but selecting german keyboard and language instead of default english[21:00:17] <tsoome> sounds like locale related issue, yep[21:16:30] <richlowe> tsoome: that's the locale bug I just talked about.[21:16:45] <tsoome> figured, yep[21:16:49] <richlowe> gea: LC_ALL=en_US.UTF-8 <run the script> will probably work.[21:16:57] <richlowe> assuming you don't nee the translations, etc.[21:23:21] <gea> if you know about, using english with current dev version is ok[21:24:10] *** melliott has quit IRC[21:30:32] *** axisys has quit IRC[21:34:57] *** flyz has quit IRC[21:35:27] *** axisys has joined #openindiana[21:38:08] *** xmikus01 has left #openindiana[21:43:17] *** tsoome has quit IRC[21:43:17] *** tsoome_ is now known as tsoome[21:44:15] *** jedi4ever has quit IRC[21:45:55] *** Norrin has quit IRC[21:46:51] *** tsoome1 has joined #openindiana[21:49:24] * RoyK is scanning through old xkcd strips http://xkcd.com/732/[21:53:39] *** Norrin has joined #openindiana[21:54:28] *** tsoome has quit IRC[21:54:28] *** tsoome1 is now known as tsoome[21:59:34] *** kdavy_ has quit IRC[22:01:22] *** kdavy_ has joined #openindiana[22:23:05] *** sudosu has quit IRC[22:23:24] <mnaser> is there a way to get the operating system root zpool (or where it is installed?)[22:23:40] <tsoome> ?[22:24:48] <tsoome> this question does not make much sense for me at least:)[22:24:48] <lennard> I don't think we know what you mean :P[22:25:11] <bdrewery> agreed[22:25:19] <alanc> you mean a way to print out the name of the zpool on which the OS is installed, if you didn't go with the default rpool?[22:25:57] <richlowe> nawk '$2 == "/" { sub("/.*", "", $1); print $1 }' /etc/mnttab[22:25:58] <alanc> "zfs list /" would list the dataset, not sure the best way to get the pool from that[22:26:00] <lennard> wow, you have good parsing skills :)[22:26:36] <alanc> or I happen to have discussed a similar question recently, and guessed...[22:26:39] <richlowe> alanc: cut -d/ -f1[22:26:54] <richlowe> alanc: now I'm mildly ashamed I went via mnttab :)[22:26:55] <lennard> just take the praise ;)[22:28:28] <mnaser> heh, yeah, i ended up with this: zfs list -H -o name / | cut -d'/' -f1[22:43:09] *** spanglywires has joined #openindiana[22:45:11] *** skeeziks has quit IRC[22:52:55] *** skeeziks has joined #openindiana[23:07:40] *** solar-star has joined #openindiana[23:11:13] *** smrt has quit IRC[23:11:31] *** smrt has joined #openindiana[23:18:17] *** pothos has quit IRC[23:18:35] *** pothos_ has joined #openindiana[23:19:01] *** pothos_ is now known as pothos[23:23:01] <DontKnwMuch> hi :) I have an interesting problem, my CPU usage in top is increasing gradually approx 3% per day of uptime. When it reached 30 or so, OI hung. After a reboot CPU usage is very low, but after 30 or so hours it is 7% already. And it is in kernel.[23:23:07] <DontKnwMuch> I run /opt/DTT/hotkernel for some time, and here are the results: http://www.pastie.org/2245076[23:23:17] *** raichoo has quit IRC[23:23:23] <DontKnwMuch> Does this look normal?[23:25:30] <DontKnwMuch> prstat -s cpu -n 5 shows: http://www.pastie.org/2241576[23:25:48] <DontKnwMuch> mpstat does have some cpu usage: http://www.pastie.org/2241586[23:26:54] <richlowe> that looks exceptionally idle[23:27:13] <dkeav> uhh huh[23:27:36] <DontKnwMuch> mpstat does show some usage on cpu 0[23:27:55] <Triskelios> the CPU power state changes take up a bit more time than expected[23:29:07] <DontKnwMuch> it does seem a liitle strange. I disabled the power management with svcadm disable power[23:29:12] <DontKnwMuch> but it did nothing[23:29:49] <DontKnwMuch> it is consuming more and more by the hour[23:30:48] <DontKnwMuch> at reboot, kernel usage is very low, lik 0.3 or so, and after two days its 30%, more it did not survive. I had it hang, not reboot nothing, just hang[23:30:53] <Triskelios> set "cpupm enable poll-mode" in power.conf, restart power[23:33:28] <DontKnwMuch> ok. I did that.[23:35:30] *** wonslung_ has quit IRC[23:36:15] <DontKnwMuch> Sampling... Hit Ctrl-C to end. ^C dtrace: couldn't stop tracing: Invalid argument[23:37:19] <DontKnwMuch> I think it will crash again... argh...[23:38:14] <DontKnwMuch> it crashed...[23:38:20] <DontKnwMuch> hung actually[23:39:37] <DontKnwMuch> its not responding on the console...[23:40:09] *** hsp has quit IRC[23:40:49] <DontKnwMuch> not to pings, nothing. The /opt/DTT/hotkernel hung it after I run it the second time to see what happens[23:44:12] <DontKnwMuch> will try to disable all the not needed stuff in bios[23:46:14] <DontKnwMuch> perhaps EIST, CPU C3/C6 P-state and stuff to off?[23:46:42] *** Transformer has joined #openindiana[23:48:29] <DontKnwMuch> wrf is WHEA support.. hm...[23:50:17] *** Transformer has quit IRC[23:51:13] <dkeav> AMD proc?[23:51:44] <DontKnwMuch> no, intel[23:51:53] <dkeav> i believe WHEA is for windows server[23:52:08] <dkeav> "windows hardware error architecture"[23:52:24] *** InTheWings has quit IRC[23:52:34] <DontKnwMuch> E3-1240 xeon, I disabled it, I also diabled VT-D and all power saving features in bios[23:53:38] <DontKnwMuch> NOTICE: SpeedStep support is being disabled due to errors parsing ACPI P-state objects exported by BIOS.[23:53:45] <DontKnwMuch> ok.. lets see now[23:59:22] <DontKnwMuch> now it is completely idle