Re: [expert] Disk activity
On Mon, Jul 16, 2001 at 04:33:21PM -0600, D. R. Evans wrote: just tried to access the floppy with no floppy in the drive. It happens quite frequently also when you do things like df or browse the /mnt/ directory, so the automounter tries to check if there's a disk. I guess that there must be some background process(es) that cause this to happen, because it happens in the middle of the night, too (according to the log). maybe some cron task ? it should be periodical To you have any suggestions as to what might be a reasonable limits.conf file that might do the job in this case? I have 128MB of physical memory in the machine, and another 128MB of swap. I never did this kind of experiments, so i'm not able to give you an answer. maybe 200 megs of hard limit for total virtual memory for the user should be fine. And an ancillary question, in case you happen to know the answer: if a program is run SETUID user x by user y, is the program subject to the limits for x or for y? I couldn't find the answer to that documented anywhere. i have two users: player and munehiro. I set this for player in limits.conf player hardas 4000 no more than 4 megs (yes i know it's 4096 really ;) ) of space. Switching to user player says [munehiro@quela munehiro]$ su - player [player@quela player]$ ulimit -a core file size (blocks) 0 data seg size (kbytes) unlimited file size (blocks) unlimited max locked memory (kbytes) unlimited max memory size (kbytes)unlimited open files 1024 pipe size (512 bytes) 8 stack size (kbytes) 8192 cpu time (seconds) unlimited max user processes 8191 virtual memory (kbytes) 4000 [player@quela player]$ trying to run a memory hungry program lead to an insuccess [player@quela player]$ xmms xmms: error while loading shared libraries: libm.so.6: cannot load shared object file: Cannot allocate memory [player@quela player]$ lynx lynx: error while loading shared libraries: libc.so.6: cannot load shared object file: Cannot allocate memory [player@quela player]$ but i'm allowed to run simple programs [player@quela player]$ ls Desktop/ collection1.zip homm3_darkage.zip homm3_evilgood.zip mail/ nsmail/ tmp/ GNUstep/ collection2.zip homm3_deadfor.zip images/ maps/ temp/ [player@quela player]$ wanna some loki heroes III maps? :) let's do some experiment [munehiro@quela munehiro]$ su - Password: [root@quela root]# chown player.player /bin/ls [root@quela root]# chmod u+s /bin/ls [root@quela root]# ls -la /bin/ls -rwsr-xr-x1 player player 44828 mar 30 16:13 /bin/ls* [root@quela root]# then i set the limit to a tight one (10 k). now [munehiro@quela /]$ whoami munehiro [munehiro@quela /]$ ls bin/ boot/ dev/ etc/ hd/ home/ lib/ mnt/ opt/ proc/ root/ sbin/ tmp/ usr/ var/ [munehiro@quela /]$ ls /home/munehiro ls: /home/munehiro: Permesso negato [munehiro@quela /]$ ls /home/player/ Desktop/ collection1.zip homm3_darkage.zip homm3_evilgood.zip mail/ nsmail/ tmp/ GNUstep/ collection2.zip homm3_deadfor.zip images/ maps/ temp/ [munehiro@quela /]$ ls -dl /home/player/ drwx-- 29 player player 4096 lug 17 18:40 /home/player// [munehiro@quela /]$ seems quite clear that even ls is suid player, the referred limit is the munehiro limit. I think this happens since the limits are set at login time, so when a suid program runs, the suid hosting user (player in this case) does not start any login session at all. -- *** This is an Honor virus. *** We have unfortunately been very busy lately and haven't had the time, nor in fact the expertise, to write a virus. So please take a couple of minutes to open Windows and randomly delete 10 or 12 files (including a minimum of 3 system files) and then send this email on to everyone on your mailing list.
Re: [expert] Disk activity
On Mon, Jul 16, 2001 at 03:34:36PM -0600, D. R. Evans wrote: Jul 16 09:31:10 localhost kernel: cdrom: open failed. Jul 16 09:31:10 localhost kernel: end_request: I/O error, dev 02:00 (floppy), sector 0 I don't know what those messages around 09:31 mean, but they seem to appear sporadically throughout the messages file, usually a few times per day. just tried to access the floppy with no floppy in the drive. It happens quite frequently also when you do things like df or browse the /mnt/ directory, so the automounter tries to check if there's a disk. If only I could figure out which process was causing the aberrant behaviour, I could start to do something about it. i've some idea. maybe wrong, but could be an hint. from my experience, in some cases, konqueror from kde2 starts sucking up ram. The disk activity you hear is just swapping activity. Since you are not ulimited, konqueror have the permission to eat up all cpu and ram, and when all virtual memory (physical and swap) is exausted, your machine is locked. if your machine is a critical routing machine, i suggest you not to hog it with this kind of processes (maybe i'm wrong, but the kde shipped with mandrake 7.2 was a pre-release, and so very buggy), and if you really need to use them, place some care in /etc/security/limits.conf. If you set correct limit parameters and an application starts to hog your machine, the kernel kills the application before it's too late. Another good solution is trying to upgrade kde :) -- Quando andarono sulla luna avevano un C64 come potenza di calcolo... Oggi senza un PIII 866 non arriviamo nemmeno oltre il peer. (munehiro)
Re: [expert] cdrecord always fail
On Sun, Jun 17, 2001 at 09:39:40AM -0400, civileme wrote: here it is. the mobo is a Asus P2BS, with onboard AHA2940 controller. The actual cdrw is correctly terminated, and the scsi id is unique. Try this rmmod aic7xxx modprobe aic7xxx_old Civileme it was not so simple. Since i have the master hd on the scsi chain, a simple modprobe does not resolve, but even crashes the whole linux box with an hard kernel panic (just to try the feeling having two conflicting drivers together :) ). So i started hacking on initrd.img, gunzipped it, mounted it on loopback, removed the module aic7xxx and replaced with the old module from the /lib/modules/2.4.3-20mdk tree. now the cdrw seems to work correctly. i've just burned an iso at 16x with no writing errors (just a couple of them in the start/stop unit, but i think it's quite normal since the drive is spinning up to catch the media paramters... it's not writing time so i dont' care) however, is not so clear why the ide cdrw had problems too... kernel hackers know :) many thanks for the help. I'll broadcast the solution as fast as i can. -- phalanx2 non ci crederai, ma ho visto un abaco con sotto scritto powered by linux
[expert] cdrecord always fail
i've a lot of problems with LM8 and cdrecord. Some weeks ago i bought a ATAPI Acer cdrw, and tried to record some cd's. Cdrecord always produced errors, so i thought the recorder was broken and i took back. I changed the cdrw with another acer. Tried this, same problem. Brought back, took another, same problem. After this, i brought back the last acer and took an yamaha 2100 SCSI. tried it. same problem. This time i'm quite sure it's not an hardware problem. I tried the cdrecord shipped with mandrake, and the cdrecord 1.11a03 (the last from the schilling site). I also recompiled the last 2.4.5 kernel. same problem. i also tried to change the media, and i tried 2 or 3 different media. No results. I started searching on deja. A lot of people have my problem, but no one get a valid and useful (for them and for me) reply. this is the report (with 1.9, the report with 1.11a03 is identical). Sometimes instead of fatal error i get a retryable error, but the writing fails in the same manner. [root@quela bin]# ./cdrecord -v speed=8 dev=/dev/sgb /mnt/iso.iso Cdrecord 1.9 (i686-pc-linux-gnu) Copyright (C) 1995-2000 Jörg Schilling TOC Type: 1 = CD-ROM scsidev: '/dev/sgb' devname: '/dev/sgb' scsibus: -2 target: -2 lun: -2 Linux sg driver version: 3.1.17 Using libscg version 'schily-0.1' atapi: 0 Device type: Removable CD-ROM Version: 2 Response Format: 2 Capabilities : SYNC Vendor_info: 'YAMAHA ' Identifikation : 'CRW2100S' Revision : '1.0H' Device seems to be: Generic mmc CD-RW. Using generic SCSI-3/mmc CD-R driver (mmc_cdr). Driver flags : SWABAUDIO Drive buf size : 7582848 = 7405 KB FIFO size : 4194304 = 4096 KB Track 01: data 629 MB Total size: 723 MB (71:38.58) = 322394 sectors Lout start: 723 MB (71:40/44) = 322394 sectors Current Secsize: 2048 ATIP info from disk: Indicated writing power: 4 Is not unrestricted Is not erasable Disk sub type: Medium Type A, high Beta category (A+) (3) ATIP start of lead in: -11079 (97:34/21) ATIP start of lead out: 359849 (79:59/74) Disk type:Long strategy type (Cyanine, AZO or similar) Manuf. index: 11 Manufacturer: Mitsubishi Chemical Corporation Blocks total: 359849 Blocks current: 359849 Blocks remaining: 37455 Starting to write CD/DVD at speed 8 in write mode for single session. Last chance to quit, starting real write in 1 seconds. Waiting for reader process to fill input buffer ... input buffer ready. Performing OPC... Starting new track at sector: 0 Track 01: 6 of 629 MB written (fifo 85%)../cdrecord: Input/output error. write_g1: scsi sendcmd: fatal error CDB: 2A 00 00 00 0C 5A 00 00 1F 00 write track data: error after 6475776 bytes Sense Bytes: 70 00 00 00 00 00 00 0A 00 00 00 00 00 00 00 00 00 00 Writing time: 21.307s Fixating... Fixating time: 31.199s ./cdrecord: fifo had 166 puts and 103 gets. ./cdrecord: fifo was 0 times empty and 20 times full, min fill was 84%. many thanks for the help. After this, i'll start downloading nero and seek for a windoze installation ;_; -- Zuglio c'e' qualche requisito minimo x linux? munehiro l'intelligenza
Re: [expert] cdrecord always fail
On Sun, Jun 17, 2001 at 01:47:34AM -0400, civileme wrote: Something clicked--can you also send the output of lspcidrake? [root@quela munehiro]# lspcidrake agpgart : Intel Corporation|440BX/ZX - 82443BX/ZX Host bridge unknown : Intel Corporation|440BX/ZX - 82443BX/ZX AGP bridge unknown : Intel Corporation|82371AB PIIX4 ISA unknown : Intel Corporation|82371AB PIIX4 IDE usb-uhci: Intel Corporation|82371AB PIIX4 USB unknown : Intel Corporation|82371AB PIIX4 ACPI - Bus Master IDE Controller aic7xxx : Adaptec|AHA-2940U2/W / 7890 ne2k-pci: Realtek|RTL-8029(AS) Card:NVIDIA GeForce2 DDR (generic): nVidia Corporation|NV11 here it is. the mobo is a Asus P2BS, with onboard AHA2940 controller. The actual cdrw is correctly terminated, and the scsi id is unique. -- ma i floppy da 1e7 tengono dieci milioni di byte?
Re: [expert] Problems with LM8.0 and The Gimp 1.2
On Thu, May 24, 2001 at 01:59:05PM +, Darrell Scott wrote: For the most part, I've been impressed (with KDE enhancements in particular). However, when I installed and tried out The Gimp (from the supplied rpm), I found it kept freezing on me! It would start fine, but when I selected another tool from the toolpad thingy, it didn't seem to want to respond any more. Sometimes it would, sometimes it wouldn't. The rest of the desktop remains responsive. Has anyone else witnessed this behaviour, and/or know of a remedy to get The Gimp working again? I was running The Gimp on KDE (this is all out-of-the-box-rpm setup). i had a hard work with gimp too. Very often (quite every ten minutes) gimp hangs with no response and no widget refresh. CPU utilization raises up to 100% and the only solution is a kill -9. i'm trying to made the behaviour reproducible, but with no success (for now). -- brevettate il nostro DNA, prima che qualche simpatica multinazionale abbia la vostra stessa idea.
[expert] cups accounting
hi all, maybe i'm wrong with configuration, but i cant' understand why a standard installation (from mdk rpms version 1.1.6, but i also tried the rpms from cups site, version 1.1.7) of cups on a mandrake 8.0 with an explicit PageLog set to /var/log/cups/page_log does not create the log file at all? for the access_log and error_log in the same directory i've no problem (both are written when an access or a print occur). I also tried to touch the unexistent page_log file, and check the permissions, but no result. Trying to log to syslog lead to the same. I found only a post on deja about my problem, but with no replies. any other had the same problem? the configuration file is a plain cupsd.conf from mdk8, with the PageLog option uncommented (but i think the default behaviour should be the logging action) many thanks -- Scopo del gioco e' vincere (munehiro)