[expert-it] mouse USB
Ciao. prima di tutto non hai detto che marca il trackball, poi dovresti provare ad eseguire anche queste righe. gpm -t pnp -m /dev/input/mice se non funziona elenca i device supportati. gpm -t help dopo controlla se realmente il device quello elencato con il comando cat /proc/bus/usb/devices cat /proc/bus/usb/drivers credo che in questo modo risolverai tutto.
R: [expert-it] Errori in pausa su ATAPI cdrom
-Messaggio originale-
Da: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]]Per conto di BanaSouKi
Inviato: martedi 23 gennaio 2001 19.16
A: [EMAIL PROTECTED]
Oggetto: [expert-it] Errori in "pausa" su ATAPI cdrom
Salve a tutti,
ho riscontrato errori e malfunzionamenti del driver ATAPI quando viene
messo in pausa il lettore CD: la lettura del disco riprende solo se la
pausa dura pochi secondi, altrimenti la lettura si blocca (sembra che il
bus venga resettato anche se non viene riportato dal dmseg). Il problema
si presenta con diverse configurazioni di sistema:
- Linux Suse 6.4 e RedHat6.2
- Kernel 2.2.14 2.2.16 2.2.18 2.4.0
- Pentium II + chipset Intel 440BX [*], K7 + chipset VIA VT 82C586 Apollo
- CDROM ASUS 40x e 50x.
[*] In questa configurazione il CDROM su canale IDE secondario si blocca
definitivamente dopo qualche minuto (ATAPI reset) ed egrave; necessario il
reboot. Sul canale primario slave il CDROM funziona, ma rimane il
problema della pausa.
Il kernel riporta i seguenti messaggi nel caso di ATAPI reset (blocco
del lettore):
Jan 22 09:39:46 saetta kernel: hdc: drive not ready for command
Jan 22 09:39:46 saetta kernel: hdc: status error: status=0xff { Busy }
Jan 22 09:39:46 saetta kernel: hdc: drive not ready for command
Jan 22 09:39:46 saetta kernel: hdc: ATAPI reset complete
Jan 22 09:39:46 saetta kernel: hdc: status error: status=0x7f {
DriveReady Devic
eFault SeekComplete DataRequest CorrectedError Index Error }
Jan 22 09:39:46 saetta kernel: hdc: status error: error=0x7f
Jan 22 09:39:46 saetta kernel: hdc: drive not ready for command
Jan 22 09:39:46 saetta kernel: hdc: ATAPI reset complete
Jan 22 09:39:46 saetta kernel: hdc: status error: status=0x7f {
DriveReady Devic
eFault SeekComplete DataRequest CorrectedError Index Error }
Jan 22 09:39:46 saetta kernel: hdc: status error: error=0x7f
Jan 22 09:39:46 saetta kernel: end_request: I/O error, dev 16:00 (hdc),
sector 0
mentre in caso di "errore in pausa":
hdc: packet command error: status=0x51 { DriveReady SeekComplete Error }
hdc: packet command error: error=0x54
ATAPI device hdc:
Error: Illegal request -- (Sense key=0x05)
Invalid field in command packet -- (asc=0x24, ascq=0x00)
The failed "Play Audio MSF" packet command was:
"47 00 00 35 0b 1b 35 02 47 00 00 00 "
Ci sono patch?
Saluti,
Marco
ciao
la macchina piu indicata per linux e il k7 perche non ha nessun problema
per ATA33-66-100, viene supportato a pieno il chipset VIA VT 82C586 (3
server in sede).
Vedendo gli errori molto probabilmente il cdrom e ATA66 o 100.
Dimenticavo Ti consiglio di installare La RedHat 6.2 o la 7 (Gia corretta
per ATA100) o la mandrake 7.1 - 7.2 e molto probabilmente risolverai anche i
problemi con il Mouse.u
Ciao Alessio
[expert-it] ....problemi con partizione vfat corrotta
Salve vorrei sapere se esiste un modo per fare un file system check da linux su una partizione vfat di WinMe. Mi successo che per 1 errore del disco (penso) il pc non riesce + a fare il boot con WinMe in quanto appena la testina dell'Hd va a leggere il settore di avvio della partizione in questione si blocca tutto e questo purtroppo accade anche facendo il boot con un dischetto di ripristino (floppy a:) in quanto appena il sistema fa una rilevazione della partizione primaria c: si blocca tutto. La cosa mi appare stranissima, anche perch invece linux monta la partizione come al solito senza problemi. Il risultato che linux l'unico sistema avviabile sul mio pc e vorrei un modo per correggere il problema, a dire il vero non so nemmeno di preciso se il problema la tabella delle partizioni o il settore di avvio della partizione WinMe Chi sa aiutarmi? Grazie a tutti. Luca
R: [expert-it] ....problemi con partizione vfat corrotta
Con fsck sulle partizione non montate (hda? e cosi via) fsck -f vfat /dev/hda? oppure avvia con il cdrom di winMe fai lo scandisk poi esegui fdisk /mbr vedi se si riavvia soltanto windows non avendo pi lilo. Dopodich avvia linux da da un dischetto e eseghi lilo. avrai al riavvio dimuovo la richiesta del sistema operativo con lilo -Messaggio originale- Da: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]Per conto di Merz Luca Inviato: mercoled 24 gennaio 2001 11.28 A: [EMAIL PROTECTED] Oggetto: [expert-it] problemi con partizione vfat corrotta Salve vorrei sapere se esiste un modo per fare un file system check da linux su una partizione vfat di WinMe. Mi successo che per 1 errore del disco (penso) il pc non riesce + a fare il boot con WinMe in quanto appena la testina dell'Hd va a leggere il settore di avvio della partizione in questione si blocca tutto e questo purtroppo accade anche facendo il boot con un dischetto di ripristino (floppy a:) in quanto appena il sistema fa una rilevazione della partizione primaria c: si blocca tutto. La cosa mi appare stranissima, anche perch invece linux monta la partizione come al solito senza problemi. Il risultato che linux l'unico sistema avviabile sul mio pc e vorrei un modo per correggere il problema, a dire il vero non so nemmeno di preciso se il problema la tabella delle partizioni o il settore di avvio della partizione WinMe Chi sa aiutarmi? Grazie a tutti. Luca
R: [expert-it] ....problemi con partizione vfat corrotta
...con fsck ho provato ma non supporta la vfat (infatti se dai un'occhiata alla directory /sbin/ al suo interno ci sono vari fsck.filesystem ma non fsck.vfat) La seconda soluzione non attuabile in quanto come avevo spiegato se carico qualsiasi altro sistema operativo appena prova ad individuare le partizioni esistenti si pianta propio tutto (=spegnere brutalmente xch non arriva nemmeno al prompt!) :-( ...che cavolo posso fare? Luca -Messaggio originale- Da: Alessio Roghi [mailto:[EMAIL PROTECTED]] Inviato: mercoled 24 gennaio 2001 11.48 A: [EMAIL PROTECTED] Oggetto: R: [expert-it] problemi con partizione vfat corrotta Con fsck sulle partizione non montate (hda? e cosi via) fsck -f vfat /dev/hda? oppure avvia con il cdrom di winMe fai lo scandisk poi esegui fdisk /mbr vedi se si riavvia soltanto windows non avendo pi lilo. Dopodich avvia linux da da un dischetto e eseghi lilo. avrai al riavvio dimuovo la richiesta del sistema operativo con lilo -Messaggio originale- Da: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]Per conto di Merz Luca Inviato: mercoled 24 gennaio 2001 11.28 A: [EMAIL PROTECTED] Oggetto: [expert-it] problemi con partizione vfat corrotta Salve vorrei sapere se esiste un modo per fare un file system check da linux su una partizione vfat di WinMe. Mi successo che per 1 errore del disco (penso) il pc non riesce + a fare il boot con WinMe in quanto appena la testina dell'Hd va a leggere il settore di avvio della partizione in questione si blocca tutto e questo purtroppo accade anche facendo il boot con un dischetto di ripristino (floppy a:) in quanto appena il sistema fa una rilevazione della partizione primaria c: si blocca tutto. La cosa mi appare stranissima, anche perch invece linux monta la partizione come al solito senza problemi. Il risultato che linux l'unico sistema avviabile sul mio pc e vorrei un modo per correggere il problema, a dire il vero non so nemmeno di preciso se il problema la tabella delle partizioni o il settore di avvio della partizione WinMe Chi sa aiutarmi? Grazie a tutti. Luca
R: [expert-it] ....problemi con partizione vfat corrotta
unaltra soluzione e provare a fare un altra partizione vfat con fdisk renderala attiva per dos in modo che diventi c e rirpovare a ripartire. Ti do una brutta notizia mi sa che hai preso un bel virus sul mbr. prima riprova a ridare lilo e basta e ripatire in modo riscrivi sull' mbr ciao -Messaggio originale- Da: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]Per conto di Merz Luca Inviato: mercoled 24 gennaio 2001 12.47 A: '[EMAIL PROTECTED]' Oggetto: R: [expert-it] problemi con partizione vfat corrotta ...con fsck ho provato ma non supporta la vfat (infatti se dai un'occhiata alla directory /sbin/ al suo interno ci sono vari fsck.filesystem ma non fsck.vfat) La seconda soluzione non attuabile in quanto come avevo spiegato se carico qualsiasi altro sistema operativo appena prova ad individuare le partizioni esistenti si pianta propio tutto (=spegnere brutalmente xch non arriva nemmeno al prompt!) :-( ...che cavolo posso fare? Luca -Messaggio originale- Da: Alessio Roghi [mailto:[EMAIL PROTECTED]] Inviato: mercoled 24 gennaio 2001 11.48 A: [EMAIL PROTECTED] Oggetto: R: [expert-it] problemi con partizione vfat corrotta Con fsck sulle partizione non montate (hda? e cosi via) fsck -f vfat /dev/hda? oppure avvia con il cdrom di winMe fai lo scandisk poi esegui fdisk /mbr vedi se si riavvia soltanto windows non avendo pi lilo. Dopodich avvia linux da da un dischetto e eseghi lilo. avrai al riavvio dimuovo la richiesta del sistema operativo con lilo -Messaggio originale- Da: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]Per conto di Merz Luca Inviato: mercoled 24 gennaio 2001 11.28 A: [EMAIL PROTECTED] Oggetto: [expert-it] problemi con partizione vfat corrotta Salve vorrei sapere se esiste un modo per fare un file system check da linux su una partizione vfat di WinMe. Mi successo che per 1 errore del disco (penso) il pc non riesce + a fare il boot con WinMe in quanto appena la testina dell'Hd va a leggere il settore di avvio della partizione in questione si blocca tutto e questo purtroppo accade anche facendo il boot con un dischetto di ripristino (floppy a:) in quanto appena il sistema fa una rilevazione della partizione primaria c: si blocca tutto. La cosa mi appare stranissima, anche perch invece linux monta la partizione come al solito senza problemi. Il risultato che linux l'unico sistema avviabile sul mio pc e vorrei un modo per correggere il problema, a dire il vero non so nemmeno di preciso se il problema la tabella delle partizioni o il settore di avvio della partizione WinMe Chi sa aiutarmi? Grazie a tutti. Luca
Re: [expert] Audio problem with xmms and kde. . .
[EMAIL PROTECTED] wrote: After consulting to kde.org they give my the way to solve that; just run xmms using "artsdsp xmms". It seems that xmms is not full compatible with arts, using this command in my case it runs fine, and also the kde sounds With other applications I have to do the same; for instance tuxafq, and now most of them are running I holp this help Francisco Alcaraz Murcia (Spain) Is there a similar solution for kde-1.1.2? I'm using LM 7.1, which I prefer over 7.2. I noticed that there is a plugin for xmms to play to kde-2's audioserver. Is it difficult to write something similar for kde-1.1.2's audioserver (if it hasn't yet been written)? Thanks. ---Norvell Spearman
Re: [expert] samba
It is not necessary to use security=share. Can your win9x box access shares on the win2k box ? Have you got Netbeui installed on win9x (a VERY BAD idea, it messes with smb over tcp/ip, since samba does not support netbeui). I suspect that your problems are mostly win9x setup problems. On a default mdk7.x install, all that I've needed to do (to access samba from win9x, winnt, win2k) was to change the workgroup and netbios name, and enable encrypted passwords. If you have troubles, post a copy of your /etc/smb.conf file. Buchan zm wrote: Andri Genio wrote: Hi Does anyone know why my samba server doesn't work in windows 98 but it works in windows 2000 ? does anyone has sample about setting up both in win2k and win98 ? cause in win98 the samba server show up in the workgroup but when i click it, it won't login. both machines i use the same pass and user please help me __ Do You Yahoo!? Yahoo! Auctions - Buy the things you want at great prices. http://auctions.yahoo.com/ Hi, make sure that you have the security = share line in the /etc/smb.conf and that you have all of your shares setup correctly. -- |Registered Linux User #182071-| Buchan MilneMechanical Engineer, Network Manager Cellphone * Work +27 82 472 2231 * +27 21 808 2497 Stellenbosch Automotive Engineering http://www.cae.co.za
Re: [expert] Overlapping partitions - help needed!
Mark Weaver wrote: On the other hand...diskdrake doesn't complain, nor crash, nor bail with an init error that is so non descript that it's error means next to nothing usable at all to the user. Bah...hum bug! I don't often spend this much time bashing and bitching about Windows stuff cause this just ain't the place for it. And neither is falling back to a real dumb program "hoping" that it's doing the job correctly and fixing the real problem. All that being said...as I stated before PM is not the answer for this situation. In common with many others, DiskDrake is NOT mature enough to be relied upon, period.Where did all your overlapping partitions come from, might I ask? DiskDrake?Others have reported this in these mailing lists. Pay attention. Failing that, look at http://deja.com You have PM, but don't seem to be aware that the Windows PM or the PM CD both give you the option to generate a pair of floppies. These boot DOS, have a DOS-based GUI and contain all the PM functionality except those crazy wizards. It is these you should use for Linux operations, not Windows. It seems you want to learn the hard way. That is one of your options. Wisdom does not come cheap. You will eventually concede I have given you a valuable lesson. -- Regards, Ron. [AU]
Re: [expert] How to move /usr to another partition?
[EMAIL PROTECTED] wrote: It's possible you had a fluke situation. I've done the procedure myself many times. Add a new drive, cp -a some of the old tree to it, update fstab and run with it. I've moved entire systems to new drives that way with no problems. I usually take a running system, partitiona nd format the new drive make a temp mount under /mnt and cp -a the data over. Remove the old section of tree (leaving a directory to mount to), remount the new partion in place and make everything permanent in fstab. This only applies to data partitions. If you move a bootable partition it will no longer be bootable. You must make the new partition bootable (For Linux: use chroot from another running linux to do all 3 of these: alter the device mapping in /etc/fstab and /etc/lilo.conf, run /sbin/lilo. For DOS/Windows: Use the DOS SYS.COM utility. For NT: Dunno). If you use a Boot Manager (eg xosl (http://www.xosl.org)) you must also set that up again. -- Regards, Ron. [AU]
Re: [expert] kmail bug
On Wednesday 24 January 2001 14:13, you wrote: I've only been using kmail for a few days now, and at least once/night it crashes. Is that typical behavoir? If so, any pointers on what is going on? I've actually installed the complete system 3 times as I play wiht install parameters before finally getting what I wanted on the current install, and with each install I am having the same kmail crashes so I don't think it is something with the install itself. Must be something with your particular setup. I was a KMail user a while bck, but used pine for a long time. Went back to try kmail with the 1.1.99 version with mdk7.2 and it's been reliable... What version it it... tim -- - Tim Therese Fairchild Atchafalaya Border Collies. Kuttabul, Queensland, Australia. - Email mailto:[EMAIL PROTECTED] Homepagehttp://members.nbci.com/amosf/bcs BCs4me Page http://members.nbci.com/amosf -
Re: [expert] Overlapping partitions - help needed!
Well...maybe I am being a bit Mule-ish about all this, but here's another reason why I "need" to know what happening and how to fix it manually. "If" I can get PM to read the disks from Windows, OR from DOS there is the risk that once the partitions problem is fixed things will have changed "inside" those partitions and the kernel no longer sees things as its expecting to find them, panics, and subsequently refuses to boot at all. New problem...partitions no longer overlapping, but kernel has no idea where the boot info is cause it's not where it was the last time it was read into memory. This has happened to me in the past and because of this I'm extremely hesitant to use partition magic, or for that matter diskdrake. Very like what is going to happen is that since the "data" on these partitions is backed up on another disk already I may find myself reinstalling just to get this situation resolved should it become too much of a problem for the OS to handle. Still, I'm at the same place, at least for now, I need to know "how" to fix this the hard way so that I can take care of it. I have a feeling that this isn't the last time that I'll see it. Mark On Wed, 24 Jan 2001, Ron Stodden wrote: Date: Wed, 24 Jan 2001 21:09:21 +1100 From: Ron Stodden [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: [expert] Overlapping partitions - help needed! Mark Weaver wrote: On the other hand...diskdrake doesn't complain, nor crash, nor bail with an init error that is so non descript that it's error means next to nothing usable at all to the user. Bah...hum bug! I don't often spend this much time bashing and bitching about Windows stuff cause this just ain't the place for it. And neither is falling back to a real dumb program "hoping" that it's doing the job correctly and fixing the real problem. All that being said...as I stated before PM is not the answer for this situation. In common with many others, DiskDrake is NOT mature enough to be relied upon, period.Where did all your overlapping partitions come from, might I ask? DiskDrake?Others have reported this in these mailing lists. Pay attention. Failing that, look at http://deja.com You have PM, but don't seem to be aware that the Windows PM or the PM CD both give you the option to generate a pair of floppies. These boot DOS, have a DOS-based GUI and contain all the PM functionality except those crazy wizards. It is these you should use for Linux operations, not Windows. It seems you want to learn the hard way. That is one of your options. Wisdom does not come cheap. You will eventually concede I have given you a valuable lesson.
Re: [expert] HELP - eth0 fails on boot, works on restart?
On Tuesday 23 January 2001 21:49, you wrote: Greetings all -- I need help!! I can't figure out why my networking won't start on boot, but works just fine if i use Drakeconf to restart networking. I have Mandrake 7.2 set up on a system with an ABIT KT7 motherboard, AMD 800 Mhz T-Bird, 256M, ATI video, Linksys LNE100TX eithernet adapter, and USR 33.6k modem (ISA on Comm 2 - ports there, system can't find modem? -- that's a problem for later) The linksys LNE100TX adapter has a big "Linux" on the box and a sample Turbolinux 6.0 Lite inside, but it has horrid problems with detection and installation. It uses the tulip driver and cannot apparently be detected or autoprobed during the install, but can be hand-configured with a tool as simple as linuxconf afterward. Once that is done, it should be visible on boot UNLESS you have PNP enabled in the BIOS. The detection of an ISA modem on install is tricky and depends partly on the motherboard BIOS settings. For best results just put /dev/ttyS1 as your modem in kppp Civileme I boot the system, eth0 fails, (ybind also) everything else works great, I use Drake conf to configure the network (it already has the IP and Domain values, I just go though it to get to the "restart networking" prompt) then say yes to restart and every thing works fine. I don't see why it should work fine but fail to initialize on boot. I then kill and restart smbd and nmdb and presto, Samba is up and running fine. I need to get this problem fixed, because if the system reboots when I'm gone, I am up a creek. I have included my ifcfg-eth0, excerpts from my boot log, and kernel/errors below in this plea for help. Any tips, suggestions or information you may have will be greatly appreciated. David Rankin [EMAIL PROTECTED] (936) 715-9333
Re: [expert] cert#37843 Just Hacked - Hidden Directory!
"Albert E. Whale" wrote: I'm not real proud of this information, but I am passing it along because of what I found and how the hackers have disguised their malicious use of Linux. I have Linux-Mandrake 7.1 with some enhancements towards 7.2. Today I found that several unwanted guests have been able to connect via ftp (not any more!). I also found some mysterious files 'running' on the server. I was able to detect the processes using the monitor utility (or top). However, I was UNABLE to find the processes in the ps -ax output?? I've never seen this before. Is this a new exploit? Imagine attempting to find a command called t0rntd on your computer, and not being able to detect in the Process List. After looking for that program and coming up blank I was able to kill the process, even though the Process ID was not detected in the Process Table (also a New one). In fact, the ONLY WAY I was able to detect this malicious process running was to perform a find command. It was found, in a directory called /usr/src/.puta/stachel/t0rntd Now mind you, you could not get a directory list for the tree /usr/src to display the .puta directory. It just wasn't there. You could cd into the directory, I have since renamed the directory. I am curious exactly how do you create a dot directory (i.e. .puta) so that it is invisible to the ls -la command? Additionally, how do you run a process and eliminate it from the process table? It appears that the strings within the applications found on my Web server are looking for Red Hat, FreeBsd, Suse and other systems with the wuftpd packed version 2.6.0. PLEASE REMOVE THIS PACKAGE FROM YOUR ENVIRONMENT! Any answers to my questions are appreciated. I have already contacted the FBI and I am monitoring my environmnet with a closer eye on the logfiles. -- Albert E. Whale - http://www.abs-comptech.com/aewhale.html -- ABS Computer Technology, Inc. - Computer Networking Specialists Sr. Network, Security and Systems Consultant HP Networking Openview, Royalty Class Consultant - http://forums.itrc.hp.com The Father's Rights Network - http://www.abs-comptech.com/frn/frnhome.html The Pennsylvania Parenthood Initiative - PAPI - Children need BOTH Parents - http://www.geocities.com/Heartland/4688/papi.htm Congrats You are the proud owner of a root kit! Basically it is a set of commands that get installed on you system that ignores the intruder. My advice is take that server offline and redo it completely you can not trust it again.As they probably have root and can play with you as they wish and then destroy the server or whatever it. So get the message? Fdisk and redo!!! Cheers Stefaans
[expert] Kppp conexcion ,error
Hi , y have MDK7.2 an when connect whit my ISP the conection of it make ok But i can't to leave to internet . I make a ping to my ip PPP conection and it response but any other no Wath happens ? Thanks
RE: Re: [expert] Audio problem with xmms and kde. . .
Sorry I have not idea for KDe 1!!! Francisco Alcaraz - Mensaje Original - Remitente: Norvell Spearman [EMAIL PROTECTED] Fecha: Mircoles, Enero 24, 2001 9:23 am Asunto: Re: [expert] Audio problem with xmms and kde. . . [EMAIL PROTECTED] wrote: After consulting to kde.org they give my the way to solve that; just run xmms using "artsdsp xmms". It seems that xmms is not full compatible with arts, using this command in my case it runs fine, and also the kde sounds With other applications I have to do the same; for instance tuxafq, and now most of them are running I holp this help Francisco Alcaraz Murcia (Spain) Is there a similar solution for kde-1.1.2? I'm using LM 7.1, which I prefer over 7.2. I noticed that there is a plugin for xmms to play to kde-2's audioserver. Is it difficult to write something similar for kde- 1.1.2'saudioserver (if it hasn't yet been written)? Thanks. ---Norvell Spearman
RE: [expert] RAID
First off, any SCSI card will do, though some won't do well with hot swap. Look into each one seprately and make your decision. Second, under linux, Mylex cards are the BOMB. They've had some availability issues lately though. We use DPT cards, but since Adaptec has bought DPT, we aren't expecting any more kernel updates to thier software. So we'll probably go Mylex with any new machines and when the time is right to upgrade the current boxen. We use software RAID here in our shop on a good number of machines, as well. Mainly the mirroring RAID 1 variety. This is with RH 6.2, but I don't think there should be many differences, if any, between RH and Mdk. Its almost scary how well it works. After using the RedHat GUI installer to do the basic setup, we were more or less done. It just worked. This was on an HP LPR, but we've had good success with Dell 2450s as well. We tried to do some very NASTY things to it. Pulled one of the two drives out while running. The kernel spit out some ugly warnings all over the screen (mainly due to SCSI bus errors), but the machine continued to function perfectly. We shut the machine down, pulled /dev/sda, and it booted off of sdb like nothing was wrong. Took out a completely identical unformatted drive and tossed it in hot. I created the proper fdisk partitions, and used the hotadd commands to rebuild the array. Shut down the machine, pulled the original sdb out, and booted off the freshly created drive. Again, it worked. Setup is important. Don't make one huge raid partition. Make a bunch of small /dev/md devices, one for each partition. You'll have less chance of data corruption should one drive or the other go down in some strange way. But then, that same advice applies with *any* server implementation. The long and short of it is, the software raid for scsi is VERY mature. If you recompile your kernel, there are certain things you NEED to have compiled in, and certain things that NEED to be in the inital ramdisk. Just read the docs that come with the package. I will lose no sleep at night because of the software RAID running at the office. This is by no means an endorsement on either my behalf or my employer's. My advice is to set up software raid on the machine before it enters production and play with it as we did. Find out how fault tolerent it is, and what you'll need to do do recover. Derek Stark IT / Linux Admin eSupportNow xt 8952 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Homer Shimpsian Sent: Tuesday, January 23, 2001 8:36 PM To: [EMAIL PROTECTED] Subject: [expert] RAID Can someone explain to me their RAID experiences in Linux? How would hot swapping work? Do U need specific software with a SCSI raid controller to handle this? Can anyone recommend a SCSI RAID controller for use with Linux Mandrake 7.1? Thanks for your advice. HA, Loadbalancing, redundancy, fail-over oh yeah
RE: Re: [expert] Audio problem with xmms and kde. . .
Have you tried the xmms-arts prog from the xmms site. That took care of all my issues of having to select either system sounds, or xmms operating correctly. Brian -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Wednesday, January 24, 2001 2:36 AM To: [EMAIL PROTECTED] Subject: RE: Re: [expert] Audio problem with xmms and kde. . . After consulting to kde.org they give my the way to solve that; just run xmms using "artsdsp xmms". It seems that xmms is not full compatible with arts, using this command in my case it runs fine, and also the kde sounds With other applications I have to do the same; for instance tuxafq, and now most of them are running I holp this help Francisco Alcaraz Murcia (Spain) - Mensaje Original - Remitente: Digital Wokan [EMAIL PROTECTED] Fecha: Martes, Enero 23, 2001 6:09 am Asunto: Re: [expert] Audio problem with xmms and kde. . . I'm having similar difficulties under Mandrake 7.2, when running the arts sound server. If I don't enable arts, KDE 2.1 (0115) doesn't playsounds, and if I do, XMMS can't play at all. Norvell Spearman wrote: I'm having problems with xmms and kde (the versions installed by LM 7.1). If I click on an mp3 file in kfm (after changing the appropriate application and mime settings), about 75% of the time xmms complains with: Couldn't open audio Please check that: 1. You have the correct output plugin selected 2. No other programs is [sic] blocking the soundcard 3. Your soundcard is configured correctly. The other 25% of the time, the mp3 starts playing as it should. I also get this error when playing mp3s from a playlist. The first plays, finishes, and then I get the above error message when the second mp3 tries to start. I have no other sound problems when it's all kde, so I'm wondering if kaudioserver is the culprit (e.g., it's having problems with non-kde apps). This problem also occurs on my workstation at the office with different hardware, same software. Thanks for any help with this. ---Norvell Spearman -- Digital Wokan Tribal mage of the electronics age Guerilla Linux Warrior
RE: [expert] Overlapping partitions - help needed!
Here's my two bits: I'd check your BIOS, and maybe your hard drive jumpers to see if LBA is turned on (probably is), and if you have some alternate disk layout. What brand is the drive? Model and size, too, if you could. Derek Stark IT / Linux Admin eSupportNow xt 8952 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Mark Weaver Sent: Wednesday, January 24, 2001 6:33 AM To: [EMAIL PROTECTED] Subject: Re: [expert] Overlapping partitions - help needed! Well...maybe I am being a bit Mule-ish about all this, but here's another reason why I "need" to know what happening and how to fix it manually. "If" I can get PM to read the disks from Windows, OR from DOS there is the risk that once the partitions problem is fixed things will have changed "inside" those partitions and the kernel no longer sees things as its expecting to find them, panics, and subsequently refuses to boot at all. New problem...partitions no longer overlapping, but kernel has no idea where the boot info is cause it's not where it was the last time it was read into memory. This has happened to me in the past and because of this I'm extremely hesitant to use partition magic, or for that matter diskdrake. Very like what is going to happen is that since the "data" on these partitions is backed up on another disk already I may find myself reinstalling just to get this situation resolved should it become too much of a problem for the OS to handle. Still, I'm at the same place, at least for now, I need to know "how" to fix this the hard way so that I can take care of it. I have a feeling that this isn't the last time that I'll see it. Mark On Wed, 24 Jan 2001, Ron Stodden wrote: Date: Wed, 24 Jan 2001 21:09:21 +1100 From: Ron Stodden [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: [expert] Overlapping partitions - help needed! Mark Weaver wrote: On the other hand...diskdrake doesn't complain, nor crash, nor bail with an init error that is so non descript that it's error means next to nothing usable at all to the user. Bah...hum bug! I don't often spend this much time bashing and bitching about Windows stuff cause this just ain't the place for it. And neither is falling back to a real dumb program "hoping" that it's doing the job correctly and fixing the real problem. All that being said...as I stated before PM is not the answer for this situation. In common with many others, DiskDrake is NOT mature enough to be relied upon, period.Where did all your overlapping partitions come from, might I ask? DiskDrake?Others have reported this in these mailing lists. Pay attention. Failing that, look at http://deja.com You have PM, but don't seem to be aware that the Windows PM or the PM CD both give you the option to generate a pair of floppies. These boot DOS, have a DOS-based GUI and contain all the PM functionality except those crazy wizards. It is these you should use for Linux operations, not Windows. It seems you want to learn the hard way. That is one of your options. Wisdom does not come cheap. You will eventually concede I have given you a valuable lesson.
RE: [expert] Overlapping partitions - help needed!
Gar. I should have asked for motherboard model as well. Derek Stark IT / Linux Admin eSupportNow xt 8952 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of D. Stark - eSN Sent: Wednesday, January 24, 2001 8:04 AM To: [EMAIL PROTECTED] Subject: RE: [expert] Overlapping partitions - help needed! Here's my two bits: I'd check your BIOS, and maybe your hard drive jumpers to see if LBA is turned on (probably is), and if you have some alternate disk layout. What brand is the drive? Model and size, too, if you could. Derek Stark IT / Linux Admin eSupportNow xt 8952 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Mark Weaver Sent: Wednesday, January 24, 2001 6:33 AM To: [EMAIL PROTECTED] Subject: Re: [expert] Overlapping partitions - help needed! Well...maybe I am being a bit Mule-ish about all this, but here's another reason why I "need" to know what happening and how to fix it manually. "If" I can get PM to read the disks from Windows, OR from DOS there is the risk that once the partitions problem is fixed things will have changed "inside" those partitions and the kernel no longer sees things as its expecting to find them, panics, and subsequently refuses to boot at all. New problem...partitions no longer overlapping, but kernel has no idea where the boot info is cause it's not where it was the last time it was read into memory. This has happened to me in the past and because of this I'm extremely hesitant to use partition magic, or for that matter diskdrake. Very like what is going to happen is that since the "data" on these partitions is backed up on another disk already I may find myself reinstalling just to get this situation resolved should it become too much of a problem for the OS to handle. Still, I'm at the same place, at least for now, I need to know "how" to fix this the hard way so that I can take care of it. I have a feeling that this isn't the last time that I'll see it. Mark On Wed, 24 Jan 2001, Ron Stodden wrote: Date: Wed, 24 Jan 2001 21:09:21 +1100 From: Ron Stodden [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: [expert] Overlapping partitions - help needed! Mark Weaver wrote: On the other hand...diskdrake doesn't complain, nor crash, nor bail with an init error that is so non descript that it's error means next to nothing usable at all to the user. Bah...hum bug! I don't often spend this much time bashing and bitching about Windows stuff cause this just ain't the place for it. And neither is falling back to a real dumb program "hoping" that it's doing the job correctly and fixing the real problem. All that being said...as I stated before PM is not the answer for this situation. In common with many others, DiskDrake is NOT mature enough to be relied upon, period.Where did all your overlapping partitions come from, might I ask? DiskDrake?Others have reported this in these mailing lists. Pay attention. Failing that, look at http://deja.com You have PM, but don't seem to be aware that the Windows PM or the PM CD both give you the option to generate a pair of floppies. These boot DOS, have a DOS-based GUI and contain all the PM functionality except those crazy wizards. It is these you should use for Linux operations, not Windows. It seems you want to learn the hard way. That is one of your options. Wisdom does not come cheap. You will eventually concede I have given you a valuable lesson.
[expert] fa 510 netgear pcmcia network adapter fails.
I posted this question on the newbie list and no one seems to have an answer, maybe one of you kind people can help.This was installed on a Toshiba Sattellite 1715 While installing Mandrake 7.2 my pcmcia card came back with this error: "* warning: insmod'ing module imm failed at /usr/bin/perl-install/modules.pm line 479".When I leave the card in at boot I get:"cardmgr[207]:executing:'modprobe' cb_enabler' " Then it hangs,hard reboot needed. lsmod shows: Module Size Used by soundcore 2800 0 (autoclean) (unused) lockd 32208 1 (autoclean) sunrpc 54640 1 (autoclean) [lockd] irda 80304 1 autofs 9456 2 (autoclean) usb-uhci 19184 0 (unused) usbcore43632 1 [usb-uhci] ds 6448 2 i82365 22928 2 pcmcia_core45984 0 [ds i82365] nls_cp437 3952 2 (autoclean) vfat9408 1 (autoclean) fat30432 1 (autoclean) [vfat] reiserfs 128592 2 modprobe tulip shows : /lib/modules/2.2.17-21mdk/net/tulip.o: init_module: Device or resource busy Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters /lib/modules/2.2.17-21mdk/net/tulip.o: insmod /lib/modules/2.2.17-21mdk/net/tulip.o failed /lib/modules/2.2.17-21mdk/net/tulip.o: insmod tulip failed I guess this means I have a resource conflict, lmao ! I am sort of a newbie, please help, I can not find any info on fixing my problem. Thanks in advance..
Re: [expert] updating modules, initrd.img, module-info
Well yes and no. Pardon my cluelessness. rpm -Uvvh kernel-2.2.17mdk-21.i586.rpm the rpm fails in post-install with some failed attempt at /sbin/installkernel -- the values for $MAPFILE and $BOOTIMAGE were wrong. the script failed and my install was left with /boot/vmlinuz--c which were results of installkernel having no idea where things were supposed to go with $1, $2, $3. This was with the /sbin/installkernel that was in the original 2.2.14mdk system of 18 months ago. I did try just setting KERNEL-VERSION=2.2.17-21mdk, BOOTIMAGE=/vmlinuz and MAPFILE=/System.map in /sbin/installkernel and running the rpm -Uvvh again. That finished the script and all was fine there with bootable System.map and vmlinuz, but now they were 2.2.16 (my last source compile from non rpm source and the process still left me with no module support other than modules that had not changed from 2.2.14, in other words no remade modules.dep perhaps? A url to a step by step updation of kernel from rpm to rpm using mandrake kernel and installkernel which covers modules and /boot requirements would be useful right about here. Or just how to get a system onto a baseline update version. A lot of the HOWTOS you see around predate rpm-style builds. thanks! -Dave Dennis -Seattle, WA On Tue, 23 Jan 2001, John Wolford wrote: Date: Tue, 23 Jan 2001 15:22:52 -0800 (PST) From: John Wolford [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: [expert] updating modules, initrd.img, module-info Hi Dave, After you're done doing # make modules modules_install bzdisk etc, do # make install and watch the screen fill up with stuff you don't have to do manually anymore! I think this answers your question, pardon me if it doesn't. -John --- David Dennis [EMAIL PROTECTED] wrote: hi alan, So lets say I was to rebuild this system and make a new initrd.img and new module-info -- what would i be doing to duplicate what was done with mandrake when it was installed? whats annoying is the old linux way of just installing modules with make modules_install and such, then running lilo, still isnt updating the stuff i need to have updated in /boot. i've read over some redhat documents (mandrake sprang up from redhat) and haven't lucked out into finding the answer yet. so the question was what needs to occur when one is recompiling ones' own kernel to rebuild the contents of the /boot directory. Can you point me towards more than just 'man mkinitrd' because that doesn't seem like it is all of it yet. thanks! -Dave D -Seattle, WA On Tue, 23 Jan 2001, Alan Shoemaker wrote: Date: Tue, 23 Jan 2001 11:36:38 -0800 From: Alan Shoemaker [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: [expert] updating modules, initrd.img, module-info David Dennis wrote: Problem: initrd.img and module-info point at the original mandrake install version files in /boot from a year ago (initrd.img - initrd-2.2.14-15mdk.img and module-info - module-info-2.2.14-15mdk even after a make bzImage in /usr/src/linux, a make modules and make modules_install, and a make bzlilo. running lilo didn't fix things either. So what happens now is my System.map and vmlinuz in /boot point at 2.2.17 files but initrd.img and module-info point at original install (2.2.14) files. I am not sure but this is probably leading to a host of module errors received upon bootup, and several important modules (sound card, scsi support) are now failing to load. Forgive this post if it belongs on newbie, I am a old style kernel compile person with limited experience with rpm. The problem is now that I have a system with 2.2.17 (the kernel version I compiled successfully using linus' tools) but in /boot there is still the initrd.img and module-info from when the system was installed. I don't want to have to reinstall to fix this, what must be done instead? (note I do not have an install CD, this was a from-the-net install. -Dave Dennis -Seattle WA [EMAIL PROTECTED] Davecheck out 'man mkinitrd'. -- Alan __ Do You Yahoo!? Yahoo! Auctions - Buy the things you want at great prices. http://auctions.yahoo.com/
RE: [expert] How to move /usr to another partition?
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of John J. LeMay Jr. Sent: Tuesday, January 23, 2001 11:45 PM To: [EMAIL PROTECTED] Subject: Re: [expert] How to move /usr to another partition? I tried 'cp -a' and um...tarballing with permissions etc, and i'm not sure if i tried anything else. The method I've found to work is to tar up usr, un-tar it in it's new home, make the change in /etc/fstab, reboot, and then remove the old usr (after remounting it as "oldusr" or something). "cp cannot handle the job," and if you remove the old usr before changing it's mount none of your libs will be found and just about everything will start failing immediately. --- NOT SO! "CP" with the proper options works just fine! I don't know how many dozen times I've done it myself. I did this yesterday to move my /usr partition to a new hard drive. As long as all relative links and file permissions are kept it works just great. Problems occur only when the softlinks span devices or partitions after the copy. Softlinks do not support this, so this can cause BIG problems, even when using TAR instead. The tar/untar suggestion is normally unfeesible on systems with disk space problems as you must have enough space to house the tar file you create. Piping the output does work, but why bother when you CAN "cp" it. -JMS
[expert] gnorpm
I tried to use gnorpm to remove a package and it left many traces if it laying around including it's menu item. Is command line rpm any better or will this happen with that as well. Darren
Re: [expert] How to move /usr to another partition?
Hope this helps-- if policy on list is mandrake employees only then kindly ignore :) For example - Your existing /usr is on /dev/hda5 and your new disk is /dev/hdd5 . Here is how to do it failsafe. (move /usr to new partition) assume you'd fdisk and mke2fs on /dev/hdd5. many howtos on that. # mount -t /dev/hdd5 /mnt/usr # cd / There are now two ways to do this - cool or cautious: cool: # tar cf - /usr | ( cd /mnt ; tar xvfp - ) [note - only one 'verbose v' cause you don't need it scrolling at you twice.] cautious: # tar cvf /mnt/usr.tar /usr # cd /mnt (should see usr.tar with length non 0) # tar xvfp usr.tar you probably don't need 'p' with modern tar commands as they usually default to permissions same p, but just in case use it . NOTE: the cautious way (2 steps) assumes you need double the space (one for the .tar and one for the extracted .tar . The cool way throws everything into memory then back onto the new partition. Much cooler if you get the filepaths correct. If you DONT get the filepaths correct you'll wind up with something like /mnt/usr/usr Depending on your attention to detail this is common mistake or needless worry :) Once this is confirmed edit /etc/fstab, comment out #/dev/hda5 /usr ext2 defaults 1 2 and make a new /usr right under it in /etc/fstab /dev/hdd5 /usr ext2 defaults 1 2 so your new /etc/fstab has the lines -- #/dev/hda5 /usr ext2 defaults 1 2 /dev/hdd5 /usr ext2 defaults 1 2 (and the rest of your mounts except swap and / are below /usr so it will still mount in the same order ls -alt /mnt/usr and if it looks like /usr did, reboot and rejoice. IF you run into problems you'll have to rescue boot, edit /etc/fstab back to the way it was, comment out /dev/hdd5, boot again and figure out what went wrong -- probably the path of the new /usr was such that you mounted /usr/usr instead of /usr - i've done that a couple of times. Good luck! -Dave Dennis -Seattle, WA On Tue, 23 Jan 2001, John Wolford wrote: Date: Tue, 23 Jan 2001 16:22:18 -0800 (PST) From: John Wolford [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: [expert] How to move /usr to another partition? Many a howto have i read. Here's the situation: My /usr partition has become full. I would like to simply create a lager partition and then copy over all the files from /usr and make the necessary changes in fstab and remount. I tried 'cp -a' and um...tarballing with permissions etc, and i'm not sure if i tried anything else. (Why? Because i had this problem a while ago and finally i gave up and reinstalled with a larger partition. This of course seems absurd but i was pressed for time and couldn't get any help from my local linux gurus) I would be "successful" but when i would boot up and log in, almost any program i ran from /usr would cause a segmentation fault and result in a core dump. I tried reinstalling the offending packages but that was a very limited solution and didn't always work (that gcc or make wouldn't work didn't make things any easier). I looked through for any symbolic link problems and i didn't have a lot (or any, that i can remember) dead links. I'm wondering if there is somewhere a reference to /dev/hdx (x=a1,b5,c, etc) that also needs to be changed. By the way, i also considered maintaining a /usr2 directory (like in, oh, say, Windows...) but one of my local linux-gurus advised me that doing that would result in an administrative nightmare. Any insight would be appreciated. Thanks, John __ Do You Yahoo!? Yahoo! Auctions - Buy the things you want at great prices. http://auctions.yahoo.com/
Re: [expert] How to move /usr to another partition?
even cooler..? :) cd to /mnt, then do tar cf - -C / usr | tar xvpf - :o) On Wed, 24 Jan 2001 06:47:21 -0800 (PST) David Dennis [EMAIL PROTECTED] wrote: cool: # tar cf - /usr | ( cd /mnt ; tar xvfp - ) [note - only one 'verbose v' cause you don't need it scrolling at you twice.]
RE: [expert] How to move /usr to another partition?
Did you mean to put the space bewteen / and usr? ie: tar cf - -C / usr | tar xvpf - ^ right there? Derek Stark IT / Linux Admin eSupportNow xt 8952 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Pop Qvarnstrm Sent: Wednesday, January 24, 2001 9:55 AM To: [EMAIL PROTECTED] Subject: Re: [expert] How to move /usr to another partition? even cooler..? :) cd to /mnt, then do tar cf - -C / usr | tar xvpf - :o) On Wed, 24 Jan 2001 06:47:21 -0800 (PST) David Dennis [EMAIL PROTECTED] wrote: cool: # tar cf - /usr | ( cd /mnt ; tar xvfp - ) [note - only one 'verbose v' cause you don't need it scrolling at you twice.]
Re: [expert] How to move /usr to another partition?
Ooops, getting tired. tar cf - -C /usr . | tar xvpf - should do better (or we'll get the /usr/usr when mounting...). Still cool though :o) On Wed, 24 Jan 2001 09:49:35 -0500 "D. Stark - eSN" [EMAIL PROTECTED] wrote: Did you mean to put the space bewteen / and usr? ie: tar cf - -C / usr | tar xvpf - ^ right there? Derek Stark IT / Linux Admin eSupportNow xt 8952 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Pop Qvarnstrm Sent: Wednesday, January 24, 2001 9:55 AM To: [EMAIL PROTECTED] Subject: Re: [expert] How to move /usr to another partition? even cooler..? :) cd to /mnt, then do tar cf - -C / usr | tar xvpf - :o) On Wed, 24 Jan 2001 06:47:21 -0800 (PST) David Dennis [EMAIL PROTECTED] wrote: cool: # tar cf - /usr | ( cd /mnt ; tar xvfp - ) [note - only one 'verbose v' cause you don't need it scrolling at you twice.]
Re: [expert] Kppp conexcion ,error
good morning, you need to configure your ISP dns ip: edit /etc/resolv.conf and add to it. good luck! Clovis Sena On 24 Jan 2001, at 9:23, Maximo Monsalvo wrote: Hi , y have MDK7.2 an when connect whit my ISP the conection of it make ok But i can't to leave to internet . I make a ping to my ip PPP conection and it response but any other no Wath happens ? Thanks
RE: [expert] Kppp conexcion ,error
that is configured -Mensaje original- De: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Enviado el: mircoles 24 de enero de 2001 11:10 Para: '[EMAIL PROTECTED]'; [EMAIL PROTECTED] Asunto: Re: [expert] Kppp conexcion ,error good morning, you need to configure your ISP dns ip: edit /etc/resolv.conf and add to it. good luck! Clovis Sena On 24 Jan 2001, at 9:23, Maximo Monsalvo wrote: Hi , y have MDK7.2 an when connect whit my ISP the conection of it make ok But i can't to leave to internet . I make a ping to my ip PPP conection and it response but any other no Wath happens ? Thanks
Re: [expert] gnorpm
You should not really use gnorpm, you should use rpmdrake ... rpm should work though. You should be able to fix the menus by running #update-menus Buchan "King, Darren" wrote: I tried to use gnorpm to remove a package and it left many traces if it laying around including it's menu item. Is command line rpm any better or will this happen with that as well. Darren -- |Registered Linux User #182071-| Buchan MilneMechanical Engineer, Network Manager Cellphone * Work +27 82 472 2231 * +27 21 808 2497 Stellenbosch Automotive Engineering http://www.cae.co.za
[expert] cdrecord - more problems - cups memory hogging
I had problems with cdrecord previously after reinstalling our printserver/cdwserver. Having compiled a new kernel, everything seemed ok until recently, when we've been losing CDs again due to "loss of streaming". Attaching a monitor to this machine, I see the following erorrs on the machine VM: do_try_to_free_pages failed for cdrecord Thus it seems that we are running out of memory. Running top shows that cupsd is consuming 17MB (56%) of our 32MB of memory, the next highest is top with 800k. Previously we had no problems with this machine (when it was running lpr). I guess my options are: 1)get more ram 2)use lpr I have become attached to cups, since we get install-and-print printing for linux boxes (better than NTs point-and-print!), but if I keep losing CDs, it's going to have to go ... Why is cups using so much memory, and is there anyway to reduce this ? I haven't kept up with the updates from the unsupported directory, since I have a working printserver for 40 clients on 2 printers, and don't want to fix something that isn't broken ... -- |Registered Linux User #182071-| Buchan MilneMechanical Engineer, Network Manager Cellphone * Work +27 82 472 2231 * +27 21 808 2497 Stellenbosch Automotive Engineering http://www.cae.co.za
RE: [expert] RAID IDE OK?
Can anyone relate their experiences with IDE RAID in Linux? I'm thinking of getting a Promise IDE RAID card.. The NT people seem to abhor the idea of software RAID or IDE RAID. If SCSI software RAID is so good, maybe it's not worth the $350 SCSI RAID card. On the other hand, if IDE RAID works equally well, I'd rather spend $700 on IDE then $4000 for SCSI. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of D. Stark - eSN Sent: Wednesday, January 24, 2001 4:53 AM To: [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: RE: [expert] RAID First off, any SCSI card will do, though some won't do well with hot swap. Look into each one seprately and make your decision. Second, under linux, Mylex cards are the BOMB. They've had some availability issues lately though. We use DPT cards, but since Adaptec has bought DPT, we aren't expecting any more kernel updates to thier software. So we'll probably go Mylex with any new machines and when the time is right to upgrade the current boxen. We use software RAID here in our shop on a good number of machines, as well. Mainly the mirroring RAID 1 variety. This is with RH 6.2, but I don't think there should be many differences, if any, between RH and Mdk. Its almost scary how well it works. After using the RedHat GUI installer to do the basic setup, we were more or less done. It just worked. This was on an HP LPR, but we've had good success with Dell 2450s as well. We tried to do some very NASTY things to it. Pulled one of the two drives out while running. The kernel spit out some ugly warnings all over the screen (mainly due to SCSI bus errors), but the machine continued to function perfectly. We shut the machine down, pulled /dev/sda, and it booted off of sdb like nothing was wrong. Took out a completely identical unformatted drive and tossed it in hot. I created the proper fdisk partitions, and used the hotadd commands to rebuild the array. Shut down the machine, pulled the original sdb out, and booted off the freshly created drive. Again, it worked. Setup is important. Don't make one huge raid partition. Make a bunch of small /dev/md devices, one for each partition. You'll have less chance of data corruption should one drive or the other go down in some strange way. But then, that same advice applies with *any* server implementation. The long and short of it is, the software raid for scsi is VERY mature. If you recompile your kernel, there are certain things you NEED to have compiled in, and certain things that NEED to be in the inital ramdisk. Just read the docs that come with the package. I will lose no sleep at night because of the software RAID running at the office. This is by no means an endorsement on either my behalf or my employer's. My advice is to set up software raid on the machine before it enters production and play with it as we did. Find out how fault tolerent it is, and what you'll need to do do recover. Derek Stark IT / Linux Admin eSupportNow xt 8952 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Homer Shimpsian Sent: Tuesday, January 23, 2001 8:36 PM To: [EMAIL PROTECTED] Subject: [expert] RAID Can someone explain to me their RAID experiences in Linux? How would hot swapping work? Do U need specific software with a SCSI raid controller to handle this? Can anyone recommend a SCSI RAID controller for use with Linux Mandrake 7.1? Thanks for your advice. HA, Loadbalancing, redundancy, fail-over oh yeah
[expert] For those with Menu troubles
Menudrake keeps files in each user directory under ~/.menu/ One of those files is named 'adding_by_menudrake'. This is where your edits go unless they are deletions in which case they go under their own name and show as "available applications" the next time you run menudrake. Your menu is built from /etc/menu/menu and other files in that directory. Menudrake is _supposed_ to be able to place a file in that directory, but it appears that it can happen only once. That file is merged with /etc/menu/menu for putposes of updating menus. If you merge by hand the data stored in your ~/.menu/adding_by_menudrake with the file referring to menudrake in your /etc directory, it will become available for all users immediately after the next run of update-menus. It will, that is, if you don't make a mistake in the editing. Otherwise, expect to be using a console to delete the merged file and rename the backups to their originals. I do NOT recommend anyone try to open /etc/menu/menu (or any other file in /etc/menu) with a text editor of less ability than jed or emacs, and specifically suggest you leave it alone with kedit, because text wrapping done by some editors causes a disaster. Anyway, that's how to update menus for now. For those of you with four copies of every menu item, look at /etc/menu/menu and look at other files there as well--those files are merged in making the menu--so your problem might lie there. Detecting and removing duplicates should be reasonably easy with a powerful editor. Civileme
[expert] Stop Auto_login
Please help. I can't figure out how to stop mandrake 7.2 from automatically logging me in? Even if my account is disabled, it still logs me into it! Security hole? bug? or was it meant to be like that? It is a horrible feature to have on a server!!! Can someone please help me. thanks Dany Allard
Re: [expert] Stop Auto_login
On Wed, 24 Jan 2001, dany allard wrote: I can't figure out how to stop mandrake 7.2 from automatically logging me in? Dany Allard /etc/sysconfig/autologin AUTOLOGIN=yes EXEC=/usr/X11R6/bin/startx USER=whomever Change the yes to a no That should do it! Otherwise ; rpm -e autologin-1.0.0-3mdk (or pertinent rpm) William Bouterse Talkeetna, Alaska
Re: [expert] Audio problem with xmms and kde. . .
Hi ! I'm having similar difficulties under Mandrake 7.2, when running the arts sound server. If I don't enable arts, KDE 2.1 (0115) doesn't play sounds, and if I do, XMMS can't play at all. try using the aRts Output Plugin for xmms. Works fine here since a few weeks. You can grap is somewhere on www.xmms.org... so long, Timo
RE: [expert] RAID IDE OK?
I've used IDE Raid without problems. The key to performance is the amount of buffer RAM on the controller. The more the better. SCSI Raid is indeed better, since each device is intelligent and can handle i/o requests independently. However SCSI raid is indeed very expensive. IDE hardware raid provides a less expensive option. Hardware IDE raid I find only marginally slower than regular IDE, with a larger buffer it may even seem faster. Higher raid levels are another matter altogether. AFAIK you can't really do RAID 5 with ide controllers... -JMS -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Homer Shimpsian Sent: Wednesday, January 24, 2001 11:04 AM To: [EMAIL PROTECTED] Subject: RE: [expert] RAID IDE OK? Can anyone relate their experiences with IDE RAID in Linux? I'm thinking of getting a Promise IDE RAID card.. The NT people seem to abhor the idea of software RAID or IDE RAID. If SCSI software RAID is so good, maybe it's not worth the $350 SCSI RAID card. On the other hand, if IDE RAID works equally well, I'd rather spend $700 on IDE then $4000 for SCSI. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of D. Stark - eSN Sent: Wednesday, January 24, 2001 4:53 AM To: [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: RE: [expert] RAID First off, any SCSI card will do, though some won't do well with hot swap. Look into each one seprately and make your decision. Second, under linux, Mylex cards are the BOMB. They've had some availability issues lately though. We use DPT cards, but since Adaptec has bought DPT, we aren't expecting any more kernel updates to thier software. So we'll probably go Mylex with any new machines and when the time is right to upgrade the current boxen. We use software RAID here in our shop on a good number of machines, as well. Mainly the mirroring RAID 1 variety. This is with RH 6.2, but I don't think there should be many differences, if any, between RH and Mdk. Its almost scary how well it works. After using the RedHat GUI installer to do the basic setup, we were more or less done. It just worked. This was on an HP LPR, but we've had good success with Dell 2450s as well. We tried to do some very NASTY things to it. Pulled one of the two drives out while running. The kernel spit out some ugly warnings all over the screen (mainly due to SCSI bus errors), but the machine continued to function perfectly. We shut the machine down, pulled /dev/sda, and it booted off of sdb like nothing was wrong. Took out a completely identical unformatted drive and tossed it in hot. I created the proper fdisk partitions, and used the hotadd commands to rebuild the array. Shut down the machine, pulled the original sdb out, and booted off the freshly created drive. Again, it worked. Setup is important. Don't make one huge raid partition. Make a bunch of small /dev/md devices, one for each partition. You'll have less chance of data corruption should one drive or the other go down in some strange way. But then, that same advice applies with *any* server implementation. The long and short of it is, the software raid for scsi is VERY mature. If you recompile your kernel, there are certain things you NEED to have compiled in, and certain things that NEED to be in the inital ramdisk. Just read the docs that come with the package. I will lose no sleep at night because of the software RAID running at the office. This is by no means an endorsement on either my behalf or my employer's. My advice is to set up software raid on the machine before it enters production and play with it as we did. Find out how fault tolerent it is, and what you'll need to do do recover. Derek Stark IT / Linux Admin eSupportNow xt 8952 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Homer Shimpsian Sent: Tuesday, January 23, 2001 8:36 PM To: [EMAIL PROTECTED] Subject: [expert] RAID Can someone explain to me their RAID experiences in Linux? How would hot swapping work? Do U need specific software with a SCSI raid controller to handle this? Can anyone recommend a SCSI RAID controller for use with Linux Mandrake 7.1? Thanks for your advice. HA, Loadbalancing, redundancy, fail-over oh yeah
Re: [expert] su in X gives xalf library error
On Tue Jan 23, 2001 at 11:23:43PM -0600, ken lierman wrote: Actually, mine started working after i upgrade xalf (0.4-4.1) AND i loged out and logged back in. Maybe that version fixes the problem. I think that was the only other thing that changed ?? got me Glad you got it working. Yes, the updated xalf 0.4-4.1 was to fix the problem you were describing. Logging out and back in shouldn't have been necessary but... sometimes this stuff works a little strange... =) -- [EMAIL PROTECTED], OpenPGP key available on www.keyserver.net 1024D/FE6F2AFD 88D8 0D23 8D4B 3407 5BD7 66F9 2043 D0E5 FE6F 2AFD - Danen Consulting Serviceswww.danen.net, www.freezer-burn.org - MandrakeSoft, Inc. Security www.linux-mandrake.com Current Linux uptime: 3 days 23 hours 14 minutes.
Re: [expert] su Problems in GNOME
On Mon Jan 22, 2001 at 06:48:39PM -0800, Baker Phil wrote: Just noticed a problem on two PCs, both running LM 7.2. One PC is at work, one at home. The two boxes are do not share a LAN connection. In the GNOME Terminal, typing "su" at a non privaledged prompt, I get the following error: $ su su: error in loading shared libraries: libxalflaunch.so.0: cannot open shared object file: No such file or directory $ You need to update xalf from updates. I take it you've updated glibc, but not xalf. Get the new xalf rpm (0.4-4.1) and upgrade it and it will work again. And yes, this only happens with the gnome-terminal... the others (not gnome-specific, like Eterm) don't have this problem. -- [EMAIL PROTECTED], OpenPGP key available on www.keyserver.net 1024D/FE6F2AFD 88D8 0D23 8D4B 3407 5BD7 66F9 2043 D0E5 FE6F 2AFD - Danen Consulting Serviceswww.danen.net, www.freezer-burn.org - MandrakeSoft, Inc. Security www.linux-mandrake.com Current Linux uptime: 3 days 23 hours 15 minutes.
Re: [expert] Real Player in Mozilla]
Sujeet Bhatt wrote: If anyone has managed to make Real Player work with Mozilla under Mandrake 7.2, please let me know how you did it. This is for the Mozilla 0.7 release built from the source tarball, and RealPlayer 7 (RealPlayer-7_0-1_i386_rpm). I had to manually add the mime description to the Helper Applications dialog under Edit-Preferences. Mozilla seems to be pretty picky about filling in ALL the fields in the New Type dialog and seems to ignore the .mime.types and .mailcap files in your $HOME. Description: RealAudio Extensions: ra rm ram MIME type: audio/x-pn-realaudio Application: /usr/lib/RealPlayer7/realplay I didn't mess with the plugins at all. duane
[expert] Security Policies
Hello there, Anyone knows of a good starting point to define security policies for a company?. I am thinking of security in the network, backup, passwords, maintence etc. I wood be glad for some pointers to the work of others, could be examples from other companies. Anyone have experience with this? Mads Rasmussen / CiT Systems www.cit.com.br
Re: [expert] Real Player in Mozilla - what about flash?
Praedor Tempus wrote: What I want to know is how to make the shockwave flash plugin work with mozilla (or konqueror for that matter). I have the latest shockwave flash netscape plugin in /usr/lib/netscape/plugins and it is, apparently, invisable to konqueror. I also copied it to /usr/local/mozilla/plugins but it doesn't work. I'm using mozilla 0.7 built from the source and have shockwave 4 r12 running. Did you copy both ShockwaveFlash.class and libflashplayer.so to the mozilla plugins directory? If you build mozilla from source it uses its src/mozilla/dist/bin/plugins directory instead of any pre-existing dirs. duane
Re: [expert] Overlapping partitions - help needed!
On Wednesday 24 January 2001 04:48, you wrote: Ok... First of all, in order for me to use Partition Magic, which I have on my windows side and hardly ever use. I'd rather use Linux...anyway, PM won't do anything with ReiserFS. Thus, in order to use PM I would have to wipe the drive, start over and make the file system ext2 in order to be able to use PM. I really don't want to do that. I've spent a lot of time setting this machine up. Another real good reason for not using PM is because I've left the windows world behind because I just plain got sick and tired or using a GUI for anything and everything. All that was doing was letting something else do the real work for me and I was learning "how" to fix a problem cause I had no idea what the program was doing while it was fixing something. Ergo, one of the biggest reasons I became a full time Linux user. And I happen to like diskdrake. It's a very powerful tool and used properly can do just as good a job as PM. And it's free! Not to mention I save the Window parition with it the other night. Not that I need or windows, but there is still a great deal of data still being stored on that system. If I'm going to "fix" this problem then I want to know why it's happening and how to deal with it without allowing a GUI to do the thinking and work for me. I like being the one making the decisions whether for good or for bad. I like choices. Therefore, I really need to know whats going on here and just how to go about fixing it. O yeah...and one more thing. Even when I did have the ext2 FS in the drive PM would bail on me each and every time without fail when I would start it and it only behaves this way when the Linux partitions are overlapped. It ain't a doin me much good that-a way. Seems the program only works for me when there "isn't" a problem on the drive that it was designed to handle. Darn! can't use it again... On the other hand...diskdrake doesn't complain, nor crash, nor bail with an init error that is so non descript that it's error means next to nothing usable at all to the user. Bah...hum bug! I don't often spend this much time bashing and bitching about Windows stuff cause this just ain't the place for it. And neither is falling back to a real dumb program "hoping" that it's doing the job correctly and fixing the real problem. All that being said...as I stated before PM is not the answer for this situation. Mark Ok, if you are done with the rant, how 'bout some data? Use what I suggested or use /sbin/fdisk -l /dev/hdx where x is the drive letter, but let's see what you are discussing. Civileme confers on Ron the "PM Salesman of the Weak" award j/k of course ;-}
Re: [expert] Real Player in Mozilla]
duane voth wrote: This is for the Mozilla 0.7 release built from the source tarball, and RealPlayer 7 (RealPlayer-7_0-1_i386_rpm). I had to manually add the mime description to the Helper Applications dialog under Edit-Preferences. Mozilla seems to be pretty picky about filling in ALL the fields in the New Type dialog and seems to ignore the .mime.types and .mailcap files in your $HOME. Description: RealAudio Extensions: ra rm ram MIME type: audio/x-pn-realaudio Application: /usr/lib/RealPlayer7/realplay I didn't mess with the plugins at all. duane Thanks for the reply. I have already done what you suggest but have had no joy so far. I am using Mozilla 0.7 and RealPlayer8. Sujeet
