Re: specifying -s flag for openntpd in rc.conf does not work
2012/4/4 Andrey N. Oktyabrski a...@bestmx.ru: P.S. To the mail list administrator: DragonFly mail lists have not the Reply-To header. Is it possible to add it in the mail lists manager settings? I forget to Reply all regularly and press Reply instead :-( seconding -- wbr, |\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`' -. ;-;;,_ McLone at GMail dot com |,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: FAILURE - READ_DMA48 status=51READY,DSC,ERROR error=10NID_NOT_FOUND LBA=441504576
On Tue, Apr 3, 2012 at 17:30, Siju George sgeorge@gmail.com wrote: On Mon, Apr 2, 2012 at 8:45 PM, McLone mcl...@gmail.com wrote: looks like something is wrong with an interface, I have plugged in the Disk to another SATA port. I guess by interfce you ment the SATA port? i meant a port + PHY + SATA chip, as a whole. Do I need to use a newer version of smartctl? well you can try and see if it fixes those spooky numbers. Problematic LBAs are always assorted, or grouped? I didn't get this. Could you please explain? :-) the logical block addresses you see in error logs may be random, or problematic LBAs can fall into one or more ranges. For example it can be 520 to 535 or so, or from, say, 7100 to the very end of disk... What LBAs generate errors in Your observations? Did You tried a simple dd sequential read/write? Could you please let me know some test I can do to check for delays? I guess i've mentioned this earlier: under winblows You can use HDTach (it shows nice graphs), alternatively You can boot from MHDD iso and use that (F4 IIRC, and then You can command it to skip fwd/back with arrow keys to reassure that the delay is right there), alternatively You can scribble some script to either SIGINFO dd and plot dd's stderr, or to fork dd (in a single user system) to do one chunk at a time and measure that time, but i guess it'll be just too crude :-) -- wbr,|\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`'-. ;-;;,_ McLone at GMail dot com|,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: ed0: NIC memory corrupt - invalid packet length 38018
On Tue, Apr 3, 2012 at 12:19, Siju George sgeorge@gmail.com wrote: On Tue, Apr 3, 2012 at 12:59 PM, Siju George sgeorge@gmail.com wrote: ed0: NIC memory corrupt - invalid packet length 27637 omg, ed is just plain scary. It never worked stable in my experience. Never. Ditch it. over cable After attaching an additional NIC to each and re-route the mirroring through them? you can, but what for? If it will go through some switch just OK Also is there a way to bypass ssh in case of large amount of Data? something like that? http://membled.com/work/patches/openssh/ -- wbr, |\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`' -. ;-;;,_ McLone at GMail dot com |,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: ed0: NIC memory corrupt - invalid packet length 38018
On Tue, Apr 3, 2012 at 18:03, McLone mcl...@gmail.com wrote: Also is there a way to bypass ssh in case of large amount of Data? something like that? http://membled.com/work/patches/openssh/ oh wait, i believe this will be much better: http://www.psc.edu/networking/projects/hpn-ssh/ -- wbr, |\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`' -. ;-;;,_ McLone at GMail dot com |,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: FAILURE - READ_DMA48 status=51READY,DSC,ERROR error=10NID_NOT_FOUND LBA=441504576
On Mon, Apr 2, 2012 at 14:40, Siju George sgeorge@gmail.com wrote: ad3: WARNING - READ_DMA48 UDMA ICRC error (retrying request) LBA=522720992 ... 5 Reallocated_Sector_Ct 0x0033 100 100 036Pre-fail Always - 0 9 Power_On_Hours 0x0032 099 099 000Old_age Always - 1173 smart table looks like from pretty new disk, but too much seek errors, and these two lines are spooky: 188 Command_Timeout 0x0032 100 098 000Old_age Always - 12885098499 240 Head_Flying_Hours 0x 100 253 000Old_age Offline - 264320877331603 looks like something is wrong with an interface, not the disk media (and You say it makes no strange sounds). Plus it looks like smartctl needs some patch :-) Problematic LBAs are always assorted, or grouped? Any unusual delays? -- wbr, |\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`' -. ;-;;,_ McLone at GMail dot com |,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: FAILURE - READ_DMA48 status=51READY,DSC,ERROR error=10NID_NOT_FOUND LBA=441504576
On Fri, Mar 30, 2012 at 14:39, Siju George sgeorge@gmail.com wrote:
Is this a problem with Hard Disk?
ad3: FAILURE - READ_DMA48 status=51READY,DSC,ERROR
seems so...
error=10NID_NOT_FOUND LBA=441504576
honestly i never had it complain about NID and idunno what is NID either.
# smartctl -l selftest /dev/ad3
you better look in smart table (-a -i, look for remap events in 1st place),
and test with some low-level tool, like MHDD or dd{_resque}.
What was the disk's history?
Was it damaged by external force?
Was it overheated? How many time is on spindle?
How many parks-unparks?
Any strange sounds?
Is read/write graph without strange apexes? (MHDD shows unusual delays,
too, but hdtach can show nice graphs; one can even SIGINFO dd and graph that)
--
wbr, |\ _,,,---,,_ dog bless ya!
` Zzz /,`.-'`' -. ;-;;,_
McLone at GMail dot com |,4- ) )-,_. ,\ ( `'-'
net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: FAILURE - READ_DMA48 status=51READY,DSC,ERROR error=10NID_NOT_FOUND LBA=441504576
On Fri, Mar 30, 2012 at 20:45, Siju George sgeorge@gmail.com wrote:
you better look in smart table (-a -i, look for remap events in 1st place),
Could you please tell me how to do this?
smartctl -s on dev smartctl -a -i dev i guess...
first command is unnecessary now-a-days, was mandatory
with old disks/BIOSes that didn't enable SMART on boot.
Was it overheated?
The server is in A/C room is there any way to find this ?
it really depends on disk, some disks have thermo sensors,
so you can even get ASCII graphs from latest smartctl.
How many time is on spindle?
How many parks-unparks?
I will need help here to find this :)
smartctl -a -i will show that too, i guess
see also here:
http://lists.freebsd.org/pipermail/freebsd-questions/2011-July/232257.html
and here (scroll down to 3rt occurrance of NID):
http://wiki.freebsd.org/JeremyChadwick/ATA_issues_and_troubleshooting
I'd also test it with some other kernel, and with
sequential then random {reads, then writes}...
What is it, btw? (the drive) Which model? Does it need FW upgrade?
--
wbr, |\ _,,,---,,_ dog bless ya!
` Zzz /,`.-'`' -. ;-;;,_
McLone at GMail dot com |,4- ) )-,_. ,\ ( `'-'
net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: Which is ideal with HAMMER? softraid or hammer volume_add
On Tue, Dec 27, 2011 at 15:41, Zenny garbytr...@gmail.com wrote: (http://leaf.dragonflybsd.org/mailarchive/commits/2009-12/msg00068.html). you need raid 0, right? i'd do volume-add if all you need is simple fileserver. In case you absolutely need striping (to split system's job between multiple spindles), i'd do natacontrol maybe, but then i'm not an expert in how hammer takes high loads itself. Both links inaccessible at the time of this writing as the dragonfly server seems to be down. seems to work for me... As always. -- wbr, |\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`' -. ;-;;,_ McLone at GMail dot com |,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
needing help deciding about upgrade to i5
hell Low. I'm having a chance to dump my E8400 for a decent price, so i need someone with one of those Intel CPUs: i5-2390T i5-2400 i5-2400S i5-2500 i5-2500S i5-2500T or any of those ending with K (unlocked multiplier) to check if CPU supports segmentation in long mode. Basically i need to know if it supports 64bit virtualbox guest. This can be checked by looking for VMX tag, or by running http://goo.gl/zcedL from vmware (win and lin binaries there). Intel's site is not crystal clear about that. Thank You. -- wbr, |\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`' -. ;-;;,_ McLone at GMail dot com |,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: needing help deciding about upgrade to i5
On Fri, Aug 26, 2011 at 1:13 PM, Sepherosa Ziehau sepher...@gmail.com wrote: I would recommend i7 instead of using i5 Life throws ENOMONEY error on that. however, if possible e3-12xx will be nice to have (mainly due to the lacking of ecc support in i7) I'm not seeing those on ark.intel.com Besides i3s do not have even VT-d. -- wbr, |\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`' -. ;-;;,_ McLone at GMail dot com |,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: How to expand HAMMER running under LVM?
On Sat, Nov 6, 2010 at 10:18 PM, Jan Lentfer jan.lent...@web.de wrote: elekktrett...@exemail.com.au schrieb: I mean some equivalent of growfs for HAMMER. ...[snip]... an LVM volume to which i add a disk to expand it, how should i grow the filesystem? via volume-add? I am not sure if there is a way to do that right now, if not, it is definitley the next logical step to have a hammer_grow or such. and don't forget about hammer_shrink ! if only i had more free time, i'd implement it... all the building blocks are in the tree... but ENOTIME :-( -- wbr, |\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`' -. ;-;;,_ McLone at GMail dot com |,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
OT: Wikipedia FS comparsion page needs update
Hell Low. I think we should conduct a slight PR move and update [1] w.r.t HAMMER. What can be said about data layout? The table at [1] has following columns: ? Block suballocation ? Variable file block size ? Extents ? Allocate-on-flush + Sparse files - Transparent compression What are those limitations for HAMMER? ? Maximum filename length ? Allowable characters in directory entries (are there exceptions?) ? Maximum pathname length ? Maximum file size Does it support ACLs, Security/MAC labels, extended attributes, alternate data streams, forks? What can be said about CoW, Block journaling or Metadata-only journaling of HAMMER? Can something be said w.r.t XIP [2]? I guess HAMMER volume still cannot be shrunk, nor can it deduplicate. [1] http://en.wikipedia.org/wiki/Comparison_of_file_systems [2] http://en.wikipedia.org/wiki/Execute_in_place -- wbr,|\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`'-. ;-;;,_ McLone at GMail dot com|,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: [HEADS UP] pkgsrc stuff
On Thu, May 6, 2010 at 5:13 AM, Justin C. Sherrill jus...@shiningsilence.com wrote: Is anyone still using DragonFly systems older than 2.4? I have old aliases for the original pkg_radd path on avalon, and if they are unused, I'd like to dump them. I have a pair of 2.2.1 boxes and i can not replace them yet. They are remote and under load. -- wbr,|\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`'-. ;-;;,_ McLone at GMail dot com|,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: Security process
On Tue, Mar 9, 2010 at 2:45 AM, Jonas Trollvik jon...@gmail.com wrote: As to blocking repeated login failures, there are such things. Doesn't pf have ip blacklisting based on certain rules built in? in Free- and OpenBSD i use this: TCP= proto tcp SSA=flags S/SA MSF=modulate state ,,, table sshlock persist ... pass in on $ext_if $TCP to $ext_addr port ssh $SSA label SSH-Limit \ $MSF (max-src-conn-rate 10/60, overload sshlock flush global) block drop in log on $ext_if from sshlock to any label SSH-Lock ... and then one can convince his cron to periodically flush that table. IIRC, DragonFly's rather old PF implementation can not handle max-src-conn-rate. It can handle max-src-states, tho. Go test it :-) -- wbr,|\ _,,,---,,_ dog bless ya! ` Zzz /,`.-'`'-. ;-;;,_ McLone at GMail dot com|,4- ) )-,_. ,\ ( `'-' net- and *BSD admin '---''(_/--' `-'\_) ...translit rawx!
Re: pkgsrc in git?
On Fri, Sep 11, 2009 at 10:58 AM, elekktrett...@exemail.com.au wrote: Now that we've been using git, do you think it may be a good idea to create a git mirror of pkgsrc? as far as i saw, NetBSD pkgsrc cvs mirrors are not reliable, and their downtime stroke me many times. pkgsrc-wip cvs seems reliable. I second on maintaining git mirror of pkgsrc, but i don't know the amount of work (and traffic) involved. I'll ponder it; then, if time/resources permit, i will mirror pkgsrc and pkgsrc-wip here in Ukraine, at least in cvs (tired of downtime, that is) -- wbr, McLone a.k.a bogon a.k.a combiner
