Back to Home / #uml / 2007 / 02 / Prev Day | Next Day
#uml IRC Logs for 2007-02-23

---Logopened Fri Feb 23 00:00:45 2007
00:07|-|ram [] has quit [Ping timeout: 480 seconds]
00:19|-|ram [] has joined #uml
00:49|-|peterz [] has quit [Ping timeout: 480 seconds]
01:07|-|fo0bar_ changed nick to fo0bar
01:14|-|peterz [] has joined #uml
01:31<fo0bar>linux-2.6 2900 root 4u REG 0,16 402653184 7969 /dev/shm/vm_file-yLrhvY (deleted)
01:32<fo0bar>doh, when did that change from /tmp...
01:38|-|mode2_ [] has joined #uml
01:43|-|mode2 [] has quit [Ping timeout: 480 seconds]
02:09|-|kos_tom [] has joined #uml
03:14|-|mode2_ [] has quit [Ping timeout: 480 seconds]
03:24|-|kos_tom [] has quit [Quit: I like core dumps]
03:49|-|AllenJB_ [] has joined #uml
04:38|-|tyler [~tyler@] has joined #uml
04:48|-|weasel [] has quit [Quit: Reconnecting]
04:48|-|weasel [weasel@2001:858:10f:6::1] has joined #uml
07:09|-|zap-zero [] has quit [Quit: brb, taufe]
07:10|-|motp [~motp@] has joined #uml
07:11|-|apic [] has joined #uml
07:24|-|tyler [~tyler@] has quit [Remote host closed the connection]
07:31|-|tyler [~tyler@] has joined #uml
07:41|-|motp [~motp@] has quit [Quit: Leaving]
08:36|-|tyler [~tyler@] has quit [Remote host closed the connection]
08:41|-|tyler [~tyler@] has joined #uml
08:44|-|Nem^ [] has quit [Remote host closed the connection]
09:34|-|jdike [] has joined #uml
09:34<jdike>Hi guys
09:52<kokoko1>habibi jdike
09:52|-|da-x [] has quit [Ping timeout: 480 seconds]
10:02|-|da-x [] has joined #uml
11:18|-|Urgleflogue [] has quit [Server closed connection]
11:19|-|Urgleflogue [] has joined #uml
11:42|-|pgstudy [] has quit [Read error: Connection reset by peer]
11:43|-|ram [] has quit [Ping timeout: 480 seconds]
11:53|-|baroni [] has joined #uml
12:10|-|ram [] has joined #uml
12:17|-|kokoko1 [~Slacker@] has quit [Quit: Leaving]
12:24|-|ousado_ [] has joined #uml
12:30|-|ousado__ [] has quit [Ping timeout: 480 seconds]
12:59|-|kokoko1 [~Slacker@] has joined #uml
12:59<kokoko1>Any hack for the day ? :)
13:02<kokoko1>guess not
13:19|-|tyler [~tyler@] has quit [Ping timeout: 480 seconds]
13:55|-|orionrobots [~danny@] has joined #uml
13:58|-|tyler [~tyler@] has joined #uml
14:10|-|mjf [] has quit [Quit: leaving]
14:11|-|vger [] has joined #uml
14:17|-|vger [] has quit [Quit: leaving]
14:17|-|vger [] has joined #uml
14:33|-|vger [] has quit [Quit: leaving]
14:33|-|vger [] has joined #uml
14:35|-|vger [] has quit []
14:35|-|vger [] has joined #uml
14:46|-|vger [] has quit [Quit: leaving]
14:46|-|vger [] has joined #uml
14:48|-|vger [] has quit []
14:48|-|vger [] has joined #uml
14:48|-|vger [] has quit []
14:50|-|vger [] has joined #uml
14:51|-|vger [] has quit []
14:51|-|vger [] has joined #uml
14:55<orionrobots>Evening all, it looks like my swap problem is back (re and it was not that fstab entry, which may have been coincidental to a real problem. To recap, I can use swapon to mount swap, but following this if I run swapon -s or top, the UML guest will hang using 100% CPU indefinitely until it is killed. Has anyone got any ideas for this?
14:57<orionrobots>I am using a BB kernel build -, and an ubuntu root fs.
14:57|-|vger [] has quit [Quit: leaving]
14:57|-|vger [] has joined #uml
14:58|-|HuK0B [~HuK0B@] has joined #uml
15:00|-|Hunger [] has quit [Server closed connection]
15:02|-|Hunger [] has joined #uml
15:09<jdike>can you get a stack trace from it?
15:09<orionrobots>Strace seems to be very uninformative.
15:09<jdike>a stack trace
15:09<jdike>gdb it and say 'bt'
15:10<orionrobots>Oh - yes.. Good plan.
15:10<jdike>what does strace say, anyway?
15:10<orionrobots>Plenty of these "munmap(0xbfaf2000, 4096) = 0" mostly.
15:11|-|SNy [] has quit [Server closed connection]
15:11|-|SNy [] has joined #uml
15:11<jdike>exactly the same address?
15:11<orionrobots>Different addresses each time.
15:11<jdike>I'd like to see a stack trace
15:12<orionrobots>Ok.. Rustling one up..
15:13|-|vger [] has quit [Quit: leaving]
15:13<jdike>also, if you could try a newer UML, that would be handy
15:13|-|vger [] has joined #uml
15:13<jdike>2.6.18 is kinda old by now
15:14<orionrobots>Ah. I suppose I was choosing one that was not too far from the kernel I use on the host box - 2.6.10 (Ubuntu Edgy). Ubuntu host, and Ubuntu guest you see. That and too lazy too build one. I am just popping that trace onto pastebin when it catches up with me..
15:14<jdike>versions don't matter
15:15<jdike>except maybe 2.4 vs 2.6
15:20<HuK0B>hmm.. I tryed to move some umls to other machine but one old uml give me something
15:20<HuK0B>Console initialized on /dev/tty0
15:20<HuK0B>Initializing software serial port version 1
15:20<HuK0B> ubda: unknown partition table
15:20<HuK0B> ubdb: unknown partition table
15:20<HuK0B>VFS: Mounted root (ext2 filesystem) readonly.
15:21<jdike>and a hang there?
15:21<HuK0B>what I missed to set
15:21<jdike>I fixed that the other day
15:21<HuK0B>but on my other machine with other host kernel work fine
15:21<jdike>orionrobots, I don't think much of that stack trace
15:21<orionrobots>jdike: how new are you recommending? 2.6.21, or something a bit older. I note that 2.6.16 is still recommended on IRC?
15:22<jdike>do you have CONFIG_DEBUG_INFO enabled?
15:22<orionrobots>jdike: A know - not a lot of debug info - no symbol table?
15:22<jdike>the newer the better
15:22<orionrobots>I downloaded a binary kernel there you see.. Lazy...
15:22<jdike>there appear to be some symbols, but not enough
15:22<jdike>can you rebuild a newer one and see if you can make it happen?
15:23<jdike>HuK0B, it's host kernel config dependent
15:23<orionrobots>jdike: Ok.
15:25<HuK0B>hmm but where can be the problem umls with ext3 work fine but this one with ext2 just hangs as I see in host kernel I put * to ext2 too :)
15:26<HuK0B>and when I try to mount image with -o loop everything is fine
15:31|-|hfb [] has joined #uml
15:33|-|hfb [] has left #uml []
15:35|-|AllenJB [~allen@] has quit [Remote host closed the connection]
15:38<jdike>try that
15:39|-|vger [] has quit [Quit: leaving]
15:40|-|vger [] has joined #uml
15:42|-|vger changed nick to mjf
15:43|-|mjf changed nick to vger
15:54<HuK0B>ok I will try it tomorow tnx
16:14|-|AllenJB_ changed nick to AllenJB
16:21|-|vger [] has quit [Quit: leaving]
16:23|-|vger [] has joined #uml
16:34<orionrobots>jdike: mm - well after configuring and building a kernel + mods myself, I cannot currently reproduce the swap problem. It is not mounting swap on startup, I am doing so manually - but that I can probably figure out. It is no longer hanging.
16:41|-|vger [] has quit [Quit: leaving]
16:42|-|mjf [] has joined #uml
16:50|-|tyler [~tyler@] has quit [Remote host closed the connection]
16:55|-|mjf [] has quit [Quit: leaving]
16:55|-|mjf [] has joined #uml
16:59|-|arun [~arun@chobie.cs.Virginia.EDU] has quit [Ping timeout: 480 seconds]
17:04<jdike>orionrobots, well, if it happens again, get a stack trace and let me know
17:04<orionrobots>Cool. UMl all seems fine at the moment - I am onto the next problem.. Oh well.
17:05<orionrobots>:) Thanks for the help
17:06<DeHackEd>reminds me, this is just a nuissance and not dangerous. only tested on linux 2.6.19 guests. swapoff on used swapspace (RAM occupancy is not an issue) results in huge CPU usage, and very slow deallocation of memory... like 1 page (4KB) per minute... space is in a file.
17:07<DeHackEd>on my system, easy. get stuff into swap, then swapoff...
17:08<orionrobots>Hmm - I wander what your largest free hole is?
17:08<DeHackEd>the UML remains responsive
17:08<orionrobots>Maybe it is having to deal with memory fragmentation depending on what kind of processes you have running and how many..
17:09<DeHackEd>system shutdown. swapoff is just a formality really.
17:09<jdike>There's a very occassional swapoff hang which happened to me once, but this sounds different
17:09<DeHackEd>let me see if I can reproduce it. I havn't booted up this UML in a week or two.
17:10<jdike>any chance you could strace the UML while it's doing this?
17:11<DeHackEd>no problem.. just give me a few minutes...
17:11<DeHackEd>this doesn't seem to be a vanilla kernel... that might have an effect...
17:13|-|arun [~arun@chobie.cs.Virginia.EDU] has joined #uml
17:19|-|HuK0B [~HuK0B@] has quit []
17:26<DeHackEd>yeah, 2.6.20 vanilla is doing it...
17:27<DeHackEd>strace shows the main thread doing a lot of polling, read/write on fd 6 ( I think it's a communication channel to the IO thread) and SIGIO stuff.
17:34<jdike>well, that would indicate it's reading stuff back in, I guess
17:34<jdike>no clue as to why it's so slow
17:35<DeHackEd>here's the stack trace: handle_syscall -> sys_swapoff -> try_to_unuse -> unuse_mm -> unuse_vma -> unuse_pte_range
17:35<DeHackEd>I guess most of it's in mm/swapfile.c
17:50<jdike>nothing surprising there
18:35|-|orionrobots [~danny@] has left #uml []
18:36|-|arun [~arun@chobie.cs.Virginia.EDU] has quit [Ping timeout: 480 seconds]
19:00|-|arun [~arun@chobie.cs.Virginia.EDU] has joined #uml
19:08|-|ousado__ [] has joined #uml
19:15|-|ousado_ [] has quit [Ping timeout: 480 seconds]
19:47|-|jdike [] has quit [Quit: Leaving]
19:51|-|ram [] has quit [Ping timeout: 480 seconds]
21:09|-|ousado_ [] has joined #uml
21:17|-|ousado__ [] has quit [Ping timeout: 480 seconds]
21:21|-|baroni [] has quit [Remote host closed the connection]
21:51|-|tarsin [] has quit [Server closed connection]
21:51|-|tarsin [] has joined #uml
22:58|-|VS_ChanLog [] has left #uml [Rotating Logs]
22:58|-|VS_ChanLog [] has joined #uml
---Logclosed Sat Feb 24 00:00:58 2007