Re: [expert] logrotate and syslog
"Jose Orlando T. Ribeiro" wrote:
>
> Hi!
>
> I have a doubt...
>
> I'm using a LM7.2 machine to store log from some Cisco routers in my network. I
> observed that the logs weren't being rotated nor compressed. So I read the man
> pages and changed the /etc/logrotate.conf, so now the Cisco log files are
> rotated and compressed.
>
> The new problem: when the logs are rotated by logrotate the syslog service
> stops... or at least the part of syslog that should take care of my Cisco's
> logs. Then I have to manually start syslog using the "service syslog restart"
> command.
>
> Has anybody had the same problem? Knows how to solve???
Mine works fine... Note that my rotate instructions are not in
/etc/logrotate.conf...
/etc/logrotate.d/syslog contains:
/home/logs/RouterLog {
rotate 5
weekly
postrotate
/usr/bin/killall -HUP syslogd
endscript
HTH,
Pierre
Want to buy your Pack or Services from MandrakeSoft?
Go to http://.mandrakestore.com
Re: [Fwd: [Fwd: Re: [expert] logrotate and syslog]]
Hi Joe! I looked at the archives, but the bug was about a "*" in the logrotate.conf files, that caused already rotated files to be re-compressed and re-rotated... I still have to "service syslog restart" manually to restart the syslog after a logrotate. I think I should have some way to put that command in the end of my logrotate.conf, but it seems that I can't put commands on it. only commands that describe how it works (rotate, compress, etc). Thanks! orlando "joe smith" wrote: > Content-Type: text/plain; charset="iso-8859-1" > From: Joe Smith <[EMAIL PROTECTED]> > Organization: Entropy Enterprises > To: "Jose Orlando T. Ribeiro" <[EMAIL PROTECTED]> > Subject: Re: [expert] logrotate and syslog > > On Wednesday 15 August 2001 15:49, you wrote: > > > Has anybody had the same problem? Knows how to solve??? > > Check the archives; there was a well-known glitch with logrotate and syslog. > > > . -- ,~~v~~,,~~v~~, ,'. .', ,'. .', === + ====== + === / ~ \ / ~ \ /\_m m_/\/\_m m_/\ .\ +--+ /. / ! [EMAIL PROTECTED] ! \ / +--+ \ `\m/ \m/' `\m/ \m/' Want to buy your Pack or Services from MandrakeSoft? Go to http://.mandrakestore.com
[expert] logrotate and syslog
Hi! I have a doubt... I'm using a LM7.2 machine to store log from some Cisco routers in my network. I observed that the logs weren't being rotated nor compressed. So I read the man pages and changed the /etc/logrotate.conf, so now the Cisco log files are rotated and compressed. The new problem: when the logs are rotated by logrotate the syslog service stops... or at least the part of syslog that should take care of my Cisco's logs. Then I have to manually start syslog using the "service syslog restart" command. Has anybody had the same problem? Knows how to solve??? TIA orlando -- ,~~v~~,,~~v~~, ,'. .', ,'. .', === + ====== + === / ~ \ / ~ \ /\_m m_/\/\_m m_/\ .\ +--+ /. / ! [EMAIL PROTECTED] ! \ / +--+ \ `\m/ \m/' `\m/ \m/' Want to buy your Pack or Services from MandrakeSoft? Go to http://.mandrakestore.com
[expert] logrotate doesn't compress logs
Hi!
I was configuring some routers to log onto my LM box. That was ok, thanks to
Pierre Fortin.
Then I realized that my logs weren´t being rotated... so was being overwhelmed
by the file sizes... I read how to configure logrotate to do this and now it
works too... but the rotated files aren't being compressed...
in the last lines of /etc/logrotate.conf I did:
...
# system-specific logs may be configured here
# rotacao dos logs do C7513FSPO01
errors root
compress
/var/log/cisco/7513-1 {
weekly
create 0644 root root
rotate 24
}
and I tried also:
# system-specific logs may be configured here
# rotacao dos logs do C7513FSPO01
errors root
/var/log/cisco/7513-1 {
weekly
compress
create 0644 root root
rotate 24
}
and tried:
# system-specific logs may be configured here
# rotacao dos logs do C7513FSPO01
errors root
/var/log/cisco/7513-1 {
weekly
create 0644 root root
compress
rotate 24
}
so... what is going on??? I know that is the correct file since any router that
I include is logged and rotated... but I don't understand why the rotated files
aren't compressed... no error message is displayed or logged... any ideas???
TIA
orlando
--
,~~v~~,,~~v~~,
,'. .', ,'. .',
=== + ====== + ===
/ ~ \ / ~ \
/\_m m_/\/\_m m_/\
.\ +--+ /.
/ ! [EMAIL PROTECTED] ! \
/ +--+ \
`\m/ \m/' `\m/ \m/'
Re: [expert] logrotate still stuck
At 03:57 PM 05/29/01, Stephen Lawrence Jr. wrote:
>Hmm, the only problem I had on 7.1 servers was in the logrotate.d/syslog
>configuration. It was set like this:
>
>
>/var/log/mail/* {
>...
>}
>
>which is wrong.
>
>The correct entrie would be:
>
>/var/log/mail/errors {
>...
>}
>
>/var/log/mail/info {
>...
>}
>
>etc.etc.etc.
>
>
>
>The '/var/log/mail/*' was causing the previously compressed and rotated logs
>to be rotated once more, causing a whole slew of rotations of rotations of
>rotations of rotations, etc..
>
>Does this help?
Thanks, but I already figured that part out. I think this problem goes a
little deeper. A corrupt file or bad memory or some such.
>Scott Taylor wrote:
>
> > Hello,
> >
> > I have been going through the archives regarding the broken logrotate
> > script and see a few that have this same problem, but still can't find the
> > fix. Maybe I missed it.
> >
> > So far, I have cleaned out the bad /var/log/mail and news
> > directories. Updated logrotate and slocate and all the rest of LMdk7.1
> > updates and still the logrotate proc never stops. It is only happening on
> > one of three 7.1 servers, so I figure something is corrupt, but can't
> > figure out what. Any body have a fix for this?
> >
> > In case this helps: after a few minutes strace on logrotate shows these
> > lines over and over like it's in a forever loop:
> > 19779 mremap(0x4013d000, 606208, 606208, MREMAP_MAYMOVE) = 0x4013d000
> > 19779 time(NULL)= 991160911
> >
> > Thanks.
> >
> > --
> > Scott Taylor
> > Systems Administrator
> > DCT Chambers Trucking Ltd.
>
>--
>Stephen Lawrence Jr. - Programmer
>[EMAIL PROTECTED]
>California Animal Health & Food Safety Laboratory System, UC Davis
>(530)-752-4614
--
Scott Taylor
Systems Administrator
DCT Chambers Trucking Ltd.
Re: [expert] logrotate still stuck
Hmm, the only problem I had on 7.1 servers was in the logrotate.d/syslog
configuration. It was set like this:
/var/log/mail/* {
...
}
which is wrong.
The correct entrie would be:
/var/log/mail/errors {
...
}
/var/log/mail/info {
...
}
etc.etc.etc.
The '/var/log/mail/*' was causing the previously compressed and rotated logs
to be rotated once more, causing a whole slew of rotations of rotations of
rotations of rotations, etc..
Does this help?
Scott Taylor wrote:
> Hello,
>
> I have been going through the archives regarding the broken logrotate
> script and see a few that have this same problem, but still can't find the
> fix. Maybe I missed it.
>
> So far, I have cleaned out the bad /var/log/mail and news
> directories. Updated logrotate and slocate and all the rest of LMdk7.1
> updates and still the logrotate proc never stops. It is only happening on
> one of three 7.1 servers, so I figure something is corrupt, but can't
> figure out what. Any body have a fix for this?
>
> In case this helps: after a few minutes strace on logrotate shows these
> lines over and over like it's in a forever loop:
> 19779 mremap(0x4013d000, 606208, 606208, MREMAP_MAYMOVE) = 0x4013d000
> 19779 time(NULL)= 991160911
>
> Thanks.
>
> --
> Scott Taylor
> Systems Administrator
> DCT Chambers Trucking Ltd.
--
Stephen Lawrence Jr. - Programmer
[EMAIL PROTECTED]
California Animal Health & Food Safety Laboratory System, UC Davis
(530)-752-4614
[expert] logrotate still stuck
Hello, I have been going through the archives regarding the broken logrotate script and see a few that have this same problem, but still can't find the fix. Maybe I missed it. So far, I have cleaned out the bad /var/log/mail and news directories. Updated logrotate and slocate and all the rest of LMdk7.1 updates and still the logrotate proc never stops. It is only happening on one of three 7.1 servers, so I figure something is corrupt, but can't figure out what. Any body have a fix for this? In case this helps: after a few minutes strace on logrotate shows these lines over and over like it's in a forever loop: 19779 mremap(0x4013d000, 606208, 606208, MREMAP_MAYMOVE) = 0x4013d000 19779 time(NULL)= 991160911 Thanks. -- Scott Taylor Systems Administrator DCT Chambers Trucking Ltd.
[expert] Logrotate problem solved
Hi Just found reason,why logrotate stuck during job. Couldn't run COMPRESS from /etc/logrotate.conf because of missing PATH in /etc/crontab Now is working suberb. Jarmo
Re: [expert] Logrotate question
On Sunday 04 March 2001 15:04, you wrote: > jarmo wrote: > > Hi > > > > I have now lost my red line how logrotate should work. > > Have one comp. running there vers.7.1 of mandrake. > > > > Have updated latest logrotate version from 7.1 updates. > > Um, can you tell me the name of the package that you used to do this > with? Hi Mark Package is logrotate-3.3-9mdk.i586.rpm jarmo PS logrotate-3.5.2 version refuses to install because of old rpm package.
Re: [expert] Logrotate question
jarmo wrote: > > Hi > > I have now lost my red line how logrotate should work. > Have one comp. running there vers.7.1 of mandrake. > > Have updated latest logrotate version from 7.1 updates. > Um, can you tell me the name of the package that you used to do this with? -- Mark "If you don't share your concepts and ideals, they end up being worthless," "Sharing is what makes them powerful."
[expert] Logrotate question
Hi I have now lost my red line how logrotate should work. Have one comp. running there vers.7.1 of mandrake. Have updated latest logrotate version from 7.1 updates. Forinstance when running cron.daily logrotate starts ok... BUT it never stops. When looking ps ax I can see ingreasing running time. When I look /var/log I can see times changed in files what logrotate should rotate. I have set options to rotate syslog and messages daily because of growing size of both files Running here radioamateur bbs/dx-cluster and because of heavy traffig everything is written to logs... So have anybody seen this kind of behaviour of logrotate? If positive,please information how to get it working properly. No need to succes,that we should upgrade to vers 7.2... Because of time limitation,the work to be done to get all working as well as it's doing at the moment,is not worth. So simple twisted modell what's wrong with logrotate... MAN doesn't explain this. Regards Jarmo [EMAIL PROTECTED]
Re: [expert] logrotate problem
On Tue Jan 30, 2001 at 03:35:49PM -0800, Tib wrote: > I've had this problem as well, with the news and mail folders. What I've > noticed is that in both of those folders, the logs are stored as 'news.crit, > mail.warn, etc' and have no .log extension - and I think this is causing > problems, because you'll end up with things like > 'mail.info.1.gz.1.gz.2.gz.2.gz.1.gz.1.gz.1.gz.etc' - this is where the mass > quantities of files comes from. The flaw in this logic is that there are a > couple log files in the /var/log/ directory that DON'T have this problem even > though they don't have a .log extension... [shrug]. Best thing to do I guess is > update the rpm as people have noted.. it's what I'm doing. :] No, the file name has absolutely nothing to do with it. Once you upgrade logrotate, you'll notice the file names are the same yet the problem is gone. The problem is with logrotate, not the logs themselves. -- [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: 4 days 7 hours 29 minutes.
Re: [expert] logrotate problem
I've had this problem as well, with the news and mail folders. What I've noticed is that in both of those folders, the logs are stored as 'news.crit, mail.warn, etc' and have no .log extension - and I think this is causing problems, because you'll end up with things like 'mail.info.1.gz.1.gz.2.gz.2.gz.1.gz.1.gz.1.gz.etc' - this is where the mass quantities of files comes from. The flaw in this logic is that there are a couple log files in the /var/log/ directory that DON'T have this problem even though they don't have a .log extension... [shrug]. Best thing to do I guess is update the rpm as people have noted.. it's what I'm doing. :] Tib On Tue, 30 Jan 2001, Laurent Duperval wrote: > Hi, > > I've noticed that whenever logrotate is run, it can take more than 2 hours > to complete. It looks like it's doing something fishy. I'm on LM 7.1 and > this is the ls -lt |head of /var/log: > > -rw-r--r--1 root root 16497976 Jan 30 13:08 debug.log > drw---2 root root 5914624 Jan 30 13:08 mail/ > -rw---1 root root91056 Jan 30 13:08 messages > -rw-r--r--1 root root91399 Jan 30 13:08 syslog > -rw-rw-r--1 root utmp 467712 Jan 30 13:07 wtmp > -rw---1 root root49106 Jan 30 13:01 cron > -rw-r--r--1 root root 504 Jan 30 13:00 daemon.log > -rw---1 root root77890 Jan 30 13:00 sudo.log > -rw---1 root root0 Jan 30 12:36 secure > > As you can see, the *directory* size of mail' is huge! I'm trying to do a ls > in that directory and it's been running for 20 minutes with no output yet. > Any ideas why this is happening? > > L > >
Re: [expert] logrotate problem
** Reply to message from Laurent Duperval <[EMAIL PROTECTED]> on Tue, 30 Jan 2001 13:40:28 -0500 (EST) > I'd just like to add that I did a > > ls -f /var/log/mail | wc > > and I got a count of more than 101000 files! I get this quite often myself. I typically blow out the directories before I go to bed and in the am I recreate them. John LeMay Jr. Senior Enterprise Consultant NJMC, LLC. [tag] I love deadlines. I like the whooshing sound they make as they fly by. - Douglas Adams
Re: [expert] logrotate problem
On 30 Jan, Daniel Woods wrote: >> > You haven't upgraded the logrotate rpm available in the 7.1 updates >> > yet, have you? Upgrade that and your logrotate problems will go away, >> >> Ok, thanks. > > And be sure to run MandrakeUpdate to get ALL the latest updates. > Or do it manually and ftp all the rpms that have been > updated since you installed LM7.1 for programs that you use. > THat's what I'm doing. One thing that scares me a bit, though, is that a warning came up in Mandrake Update saying "Don't update this package using Mandrake Update" but I wasn't given an option of sckipping that particular package (a kernel, I think). Hopefully I won't crash too hard. > Chances are that if you have not been doing that, you may have already > had your server cracked by script kiddies... especially with the RameN Crew > worm out this month. I doubt it. I'm on a laptop behind a firewall and I don't have telnet or ftp or rlogin or sshd or httpd or most of those daemons running. L -- MY EMAIL ADDRESS HAS CHANGED --> UPDATE YOUR ADDRESSBOOK Laurent Duperval "Montreal winters are an intelligence test, Netergy Networks - Java Centerand we who are here have failed it." Phone: (514) 282-8484 ext. 228 -Doug Camilli mailto:[EMAIL PROTECTED] Penguin Power!
Re: [expert] logrotate problem
> > You haven't upgraded the logrotate rpm available in the 7.1 updates > > yet, have you? Upgrade that and your logrotate problems will go away, > > Ok, thanks. And be sure to run MandrakeUpdate to get ALL the latest updates. Or do it manually and ftp all the rpms that have been updated since you installed LM7.1 for programs that you use. Chances are that if you have not been doing that, you may have already had your server cracked by script kiddies... especially with the RameN Crew worm out this month. I discovered one university library server affected by this worm because the person responsible never did any RPM updates. Thanks... Dan.
Re: [expert] logrotate problem
On 30 Jan, Vincent Danen wrote:
> On Tue Jan 30, 2001 at 01:11:20PM -0500, Laurent Duperval wrote:
>
>> I've noticed that whenever logrotate is run, it can take more than 2 hours
>> to complete. It looks like it's doing something fishy. I'm on LM 7.1 and
>> this is the ls -lt |head of /var/log:
>
> You haven't upgraded the logrotate rpm available in the 7.1 updates
> yet, have you? Upgrade that and your logrotate problems will go away,
Ok, thanks.
> but you will still need to clean out your /var/log/mail and
> /var/log/news directories. Best method might be to "rm -rf {news,mail}"
> from your /var/log directory and recreate the directories.
>
Yep, that's what I'm doing now.
L
--
MY EMAIL ADDRESS HAS CHANGED --> UPDATE YOUR ADDRESSBOOK
Laurent Duperval "Montreal winters are an intelligence test,
Netergy Networks - Java Centerand we who are here have failed it."
Phone: (514) 282-8484 ext. 228 -Doug Camilli
mailto:[EMAIL PROTECTED] Penguin Power!
Re: [expert] logrotate problem
On Tue Jan 30, 2001 at 01:11:20PM -0500, Laurent Duperval wrote:
> I've noticed that whenever logrotate is run, it can take more than 2 hours
> to complete. It looks like it's doing something fishy. I'm on LM 7.1 and
> this is the ls -lt |head of /var/log:
You haven't upgraded the logrotate rpm available in the 7.1 updates
yet, have you? Upgrade that and your logrotate problems will go away,
but you will still need to clean out your /var/log/mail and
/var/log/news directories. Best method might be to "rm -rf {news,mail}"
from your /var/log directory and recreate the directories.
--
[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 18 hours 54 minutes.
Re: [expert] logrotate problem
On 30 Jan, To: Mandrake Expert List wrote: > Hi, > > I've noticed that whenever logrotate is run, it can take more than 2 hours > to complete. It looks like it's doing something fishy. I'm on LM 7.1 and > this is the ls -lt |head of /var/log: > > -rw-r--r--1 root root 16497976 Jan 30 13:08 debug.log > drw---2 root root 5914624 Jan 30 13:08 mail/ > -rw---1 root root91056 Jan 30 13:08 messages > -rw-r--r--1 root root91399 Jan 30 13:08 syslog > -rw-rw-r--1 root utmp 467712 Jan 30 13:07 wtmp > -rw---1 root root49106 Jan 30 13:01 cron > -rw-r--r--1 root root 504 Jan 30 13:00 daemon.log > -rw---1 root root77890 Jan 30 13:00 sudo.log > -rw---1 root root0 Jan 30 12:36 secure > > As you can see, the *directory* size of mail' is huge! I'm trying to do a ls > in that directory and it's been running for 20 minutes with no output yet. > Any ideas why this is happening? > > L > I'd just like to add that I did a ls -f /var/log/mail | wc and I got a count of more than 101000 files! L -- MY EMAIL ADDRESS HAS CHANGED --> UPDATE YOUR ADDRESSBOOK Laurent Duperval "Montreal winters are an intelligence test, Netergy Networks - Java Centerand we who are here have failed it." Phone: (514) 282-8484 ext. 228 -Doug Camilli mailto:[EMAIL PROTECTED] Penguin Power!
[expert] logrotate problem
Hi, I've noticed that whenever logrotate is run, it can take more than 2 hours to complete. It looks like it's doing something fishy. I'm on LM 7.1 and this is the ls -lt |head of /var/log: -rw-r--r--1 root root 16497976 Jan 30 13:08 debug.log drw---2 root root 5914624 Jan 30 13:08 mail/ -rw---1 root root91056 Jan 30 13:08 messages -rw-r--r--1 root root91399 Jan 30 13:08 syslog -rw-rw-r--1 root utmp 467712 Jan 30 13:07 wtmp -rw---1 root root49106 Jan 30 13:01 cron -rw-r--r--1 root root 504 Jan 30 13:00 daemon.log -rw---1 root root77890 Jan 30 13:00 sudo.log -rw---1 root root0 Jan 30 12:36 secure As you can see, the *directory* size of mail' is huge! I'm trying to do a ls in that directory and it's been running for 20 minutes with no output yet. Any ideas why this is happening? L -- MY EMAIL ADDRESS HAS CHANGED --> UPDATE YOUR ADDRESSBOOK Laurent Duperval "Montreal winters are an intelligence test, Netergy Networks - Java Centerand we who are here have failed it." Phone: (514) 282-8484 ext. 228 -Doug Camilli mailto:[EMAIL PROTECTED] Penguin Power!
Re: [expert] logrotate
On Fri, 20 Oct 2000 09:24:52 -0600, Vincent Danen wrote: >Yeah, use an ftp client (ncftp, etc.) and connect to a mirror, >download the updates, then run "rpm -Fvh *.rpm" and upgrade the >packages. Very easy. Oh, thanks. I tried ftp.mandrake.com, found NOTHING, not even a text file of explanations, so I was baffled. I'll try one of its mirrors.. thanks, -turgut EgeNet Internet Services: http://www.egenet.com.tr Find all of Turkey online: http://find.egenet.com.tr Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] logrotate
On Fri, 20 Oct 2000, Turgut Kalfaoglu wrote: > I wrote a C program that would create a batch to do just that: > rm mail.err.%d.gz.%d.gz.%d.gz.%d.gz* Can you email me the source to this? I'm starting to learn C++ and C, and I can't understand how to do system calls. Thanks a bundle! -- Asheesh. Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] logrotate
On Fri Oct 20, 2000 at 03:19:00PM +0300, turgut kalfaoglu wrote: > On Thu, 19 Oct 2000 14:45:33 -0400, Bill Shirley wrote: > >rm -Rf /var/log/news /var/log/mail > >mkdir -m 755 /var/log/news /var/log/mail > > Many thanks Bill and others for helping out! > Two servers are now happy, with a cron job > to delete all files from the mail and news directories. You really don't want to do this... There is no need to make a cron job to do this if you upgrade to the latest logrotate and sysklogd from updates. > Of course, that is not a cure; does anyone know how > servers without X can use MandrakeUpdate, or how > else can I patch these servers? Yeah, use an ftp client (ncftp, etc.) and connect to a mirror, download the updates, then run "rpm -Fvh *.rpm" and upgrade the packages. Very easy. -- [EMAIL PROTECTED], OpenPGP key available on www.keyserver.net // Danen Consulting Serviceswww.danen.net, www.freezer-burn.org // MandrakeSoft, Inc. www.linux-mandrake.com 1024D/FE6F2AFD 88D8 0D23 8D4B 3407 5BD7 66F9 2043 D0E5 FE6F 2AFD Current Linux uptime: 3 days 10 hours 27 minutes. Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
RE: [expert] logrotate
On Thu, 19 Oct 2000 14:45:33 -0400, Bill Shirley wrote: >rm -Rf /var/log/news /var/log/mail >mkdir -m 755 /var/log/news /var/log/mail Many thanks Bill and others for helping out! Two servers are now happy, with a cron job to delete all files from the mail and news directories. Of course, that is not a cure; does anyone know how servers without X can use MandrakeUpdate, or how else can I patch these servers? Thanks, -turgut EgeNet Internet Services: http://www.egenet.com.tr Find all of Turkey online: http://find.egenet.com.tr Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] logrotate
> > rm [Aa]* > > rm [Bb]* > > rm [Cc]* > > I wrote a C program that would create a batch to do just that: > rm mail.err.%d.gz.%d.gz.%d.gz.%d.gz* > where each %d goes from 1 to 9.. If I reduce it by 1 %d.gz, I get the I didn't notice the front end of this thread so I'm not sure what's going on but this looks like a dandy place for a Perl script. The problem is I don't know where these files are located. But just to hint at a solution ls | myscript.pl with @files = will generate an array of file names. You can then grab them one at a time, do anything you want to change their names, and then execute the rm command. > Argument List too long. It's naturally quite inefficient, takes hours to > run. I was hoping there was a Unix utility that would delete all files > without first expanding the parameters.. This hints that you're overrunning the bounds of rm with filenames. Again, a perl script might help. If nothing else you can stick those rm [Aa]*, rm [Bb]* commands into a single shell script. No need to get fancy I should think. Sorry if I've completely misunderstood the goal. Cheers --- Larry Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] logrotate
On Thu, 19 Oct 2000, Ron Johnson, Jr. wrote: > > [root@db2 mail]# rm * > > bash: /bin/rm: Argument list too long > Divide and conquer: > rm [Aa]* > rm [Bb]* > rm [Cc]* I wrote a C program that would create a batch to do just that: rm mail.err.%d.gz.%d.gz.%d.gz.%d.gz* where each %d goes from 1 to 9.. If I reduce it by 1 %d.gz, I get the Argument List too long. It's naturally quite inefficient, takes hours to run. I was hoping there was a Unix utility that would delete all files without first expanding the parameters.. =turgut - Turgut Kalfaoglu, EgeNet Internet Services http://www.egenet.com.tr All of Turkey Online: http://find.egenet.com.tr Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
RE: [expert] logrotate
rm -Rf /var/log/news /var/log/mail mkdir -m 755 /var/log/news /var/log/mail It works, been there, done that. Bill > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED]]On Behalf Of Ron Johnson, Jr. > Sent: Thursday, October 19, 2000 11:38 AM > To: [EMAIL PROTECTED] > Subject: Re: [expert] logrotate > > > Turgut Kalfaoglu wrote: > > > > Let's assume that you have successfully ignored the patches, > > and announcements, and ended up with a system (Mandrake 7.1) > > that has over 5 MILLION files on its /var/log/mail and > /news directories. > > > > Let's also assume that when you type: > > [root@db2 mail]# rm * > > you get: > > bash: /bin/rm: Argument list too long > > > > any ideas what you can do to save your servers? :) > > > > Thank you! > > -turgut > > > > PS: 5 million files is an estimate. > > > > - > > Divide and conquer: > rm [Aa]* > rm [Bb]* > rm [Cc]* > etc., etc. > > Ron > -- > +--+ > | Ron Johnson, Jr.Home: [EMAIL PROTECTED] | > | Jefferson, LA USA WWW : [EMAIL PROTECTED] | > | http://ronandheather.dhs.org | > +--+ > > Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] logrotate
On Thu Oct 19, 2000 at 06:06:12PM +0300, Turgut Kalfaoglu wrote: > Let's assume that you have successfully ignored the patches, > and announcements, and ended up with a system (Mandrake 7.1) > that has over 5 MILLION files on its /var/log/mail and /news directories. > > Let's also assume that when you type: > [root@db2 mail]# rm * > you get: > bash: /bin/rm: Argument list too long > > any ideas what you can do to save your servers? :) Wouldn't: rm -rf /var/log/mail; mkdir /var/log/mail work? -- [EMAIL PROTECTED], OpenPGP key available on www.keyserver.net // Danen Consulting Serviceswww.danen.net, www.freezer-burn.org // MandrakeSoft, Inc. www.linux-mandrake.com 1024D/FE6F2AFD 88D8 0D23 8D4B 3407 5BD7 66F9 2043 D0E5 FE6F 2AFD Current Linux uptime: 2 days 12 hours 44 minutes. Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] logrotate
At 6:06 PM +0300 10/19/00, Turgut Kalfaoglu wrote: >Let's assume that you have successfully ignored the patches, >and announcements, and ended up with a system (Mandrake 7.1) >that has over 5 MILLION files on its /var/log/mail and /news directories. > >Let's also assume that when you type: > [root@db2 mail]# rm * >you get: > bash: /bin/rm: Argument list too long > >any ideas what you can do to save your servers? :) This should work. The syntax is for tcsh, but it should be similar in bash: # tcsh # foreach foo (*) # /bin/rm $foo # end -- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- smalltime industries bringing you the latest in interplanetary transport technology, interactive web games, dada, art, poetry and pants. www.smalltime.com .. visit early and often -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] logrotate
Turgut Kalfaoglu wrote: > > Let's assume that you have successfully ignored the patches, > and announcements, and ended up with a system (Mandrake 7.1) > that has over 5 MILLION files on its /var/log/mail and /news directories. > > Let's also assume that when you type: > [root@db2 mail]# rm * > you get: > bash: /bin/rm: Argument list too long > > any ideas what you can do to save your servers? :) > > Thank you! > -turgut > > PS: 5 million files is an estimate. > > - Divide and conquer: rm [Aa]* rm [Bb]* rm [Cc]* etc., etc. Ron -- +--+ | Ron Johnson, Jr.Home: [EMAIL PROTECTED] | | Jefferson, LA USA WWW : [EMAIL PROTECTED] | | http://ronandheather.dhs.org | +--+ Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] logrotate
Turgut Kalfaoglu said at Ò[expert] logrotateÓ.
[2000-10-19 17:06]
> Let's assume that you have successfully ignored the patches,
> and announcements, and ended up with a system (Mandrake 7.1)
> that has over 5 MILLION files on its /var/log/mail and /news directories.
>
> Let's also assume that when you type:
> [root@db2 mail]# rm *
> you get:
> bash: /bin/rm: Argument list too long
>
> any ideas what you can do to save your servers? :)
>
Try with 'xargs' (man xargs), it is just for that.
Also you could try also inside /var/log/ with find:
find . -exec rm -f {} \;
This second solution can be MUCH slower...
Keep in touch with http://mandrakeforum.com:
Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] logrotate
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Thu, 19 Oct 2000, Turgut Kalfaoglu wrote:
>
> Let's assume that you have successfully ignored the patches,
> and announcements, and ended up with a system (Mandrake 7.1)
> that has over 5 MILLION files on its /var/log/mail and /news directories.
>
> Let's also assume that when you type:
> [root@db2 mail]# rm *
> you get:
> bash: /bin/rm: Argument list too long
>
> any ideas what you can do to save your servers? :)
Use find to delete the files one at a time. This will take a while, but it'll
get the job done. Here is a sample find command that might do the trick for
/var/log/mail:
find /var/log/mail -ctime +3 -exec rm {} \;
- --
Jim Holthaus (pronunciation: HOLT house)
[EMAIL PROTECTED]
PGP Public Key at http://www.holthaus.com/jim/pgpkey.html
Learn about PGP at http://www.holthaus.com/jim/pgp.html
-BEGIN PGP SIGNATURE-
Version: PGP 6.5.1
iQA/AwUBOe8Pun3IcJJ+eC/CEQJZBACg4OBND3URjBUZEv26Wy1Q3kCLO/oAnAjm
zCERmPkgX23ixqfe3cFeWRKM
=rm6t
-END PGP SIGNATURE-
Keep in touch with http://mandrakeforum.com:
Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] logrotate
Let's assume that you have successfully ignored the patches, and announcements, and ended up with a system (Mandrake 7.1) that has over 5 MILLION files on its /var/log/mail and /news directories. Let's also assume that when you type: [root@db2 mail]# rm * you get: bash: /bin/rm: Argument list too long any ideas what you can do to save your servers? :) Thank you! -turgut PS: 5 million files is an estimate. - Turgut Kalfaoglu, EgeNet Internet Services http://www.egenet.com.tr All of Turkey Online: http://find.egenet.com.tr Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
