Re: cvs up, no space left in /tmp
On Aug 9, 2007, at 9:46 PM, Jim Razmus wrote: * Mackan <[EMAIL PROTECTED]> [070809 14:13]: Hi, # cd /usr/src && cvs -q up -rOPENBSD_4_1 -Pd can't create temporary directory /tmp/cvs-serv19063 No space left on device I think that is an issue on the server. Here's an alternative that eases the burden on the server and speeds up your local operations. Try using cvsync to get and maintain your own copy of the repository. Then do your cvs co and cvs up operations against it. If your sending back patches based on changes in your cvs tree, just be sure to #> cvs -d [EMAIL PROTECTED]:/cvs diff to create your patches. HTH, Jim Thanks, I never thought of that the "no space left" message was from the server side. It isn't obvious. Mackan
cvs up, no space left in /tmp
Hi, For many months, I've had this in root's crontab: # update src tree 0 4 * * * cd /usr/src && cvs -q up - rOPENBSD_4_1 -Pd with no problem. For about a week now I get mail from cron about no space left on device. I need some advise on how to find whats wrong. # dmesg OpenBSD 4.1-stable (GENERIC) #0: Sat Jun 9 12:54:39 CEST 2007 [EMAIL PROTECTED]:/usr/src/sys/arch/i386/compile/GENERIC [...] real mem = 1064857600 (1039900K) avail mem = 964227072 (941628K) using 4278 buffers containing 53366784 bytes (52116K) of memory # cd /usr/src && cvs -q up -rOPENBSD_4_1 -Pd can't create temporary directory /tmp/cvs-serv19063 No space left on device # df -hi Filesystem SizeUsed Avail Capacity iused ifree %iused Mounted on /dev/wd0a 147M 37.7M102M27%2208 1699012% / /dev/wd0d 148M104K140M 0% 14 19312 0% /tmp /dev/wd0e 1006M405M551M42% 13289 130965 9% /usr /dev/wd0i 1006M 59.6M896M 6%2874 141380 2% / usr/local /dev/wd0g 2.0G9.3M1.9G 0%1733 266169 1% / usr/obj /dev/wd0h 2.0G194M1.7G10% 81258 18664430% / usr/ports /dev/wd0f 2.0G630M1.3G33% 78418 18948429% / usr/src /dev/wd0j 4.9G787M3.9G16%4819 654635 1% /var /dev/wd0k 214G5.3G198G 3% 58249 28401397 0% / var/www/users
Re: Restart Apache with clean enviroment
On 6/16/07, Mackan <[EMAIL PROTECTED]> wrote: Hi list! How do you guys restart apache (apachectl stop / start) without having all the current shell variables show up in phpinfo() that is exported in the shell? Mackan On Jun 16, 2007, at 1:59 PM, Almir Karic wrote: env -i is your friend. -- almir Perfect, thank you! Mackan
Restart Apache with clean enviroment
Hi list! How do you guys restart apache (apachectl stop / start) without having all the current shell variables show up in phpinfo() that is exported in the shell? Mackan
Re: Rotate many Apache logfiles
On 15 sep 2006, at 18.57, Garance A Drosihn wrote: At 2:57 PM +0200 9/15/06, Mackan wrote: Hi! What is the preferred way of rotating Apache's logfiles? I have many virtual domains, each with its own access and error logfile. I'm using CustomLog, not TransferLog. Apache is chrooted. Adding every logfile to /etc/newsyslog.conf is one way, but hard to maintain. Is Apache's own rotatelogs program the way to go? Fwiw, the version of newsyslog in FreeBSD supports pattern-matching on the logfile names. However, it may not have some features that are in the version of newsyslog that comes with OpenBSD. Ok. If you don't want to pull that in, then maybe setup a separate newsyslog.conf file (and a second cronjob for it). That way it should be easier to use a shell script to create the appropriate entries for that conf file, without worrying that you're going to clobber any of the standard system entries. This is exactly what I plan to do. I don't want to bring in too many ports and 3rd party stuff. Thank you, and all other nice ppl on the list, for your replies. Mackan
Rotate many Apache logfiles
Hi! What is the preferred way of rotating Apache's logfiles? I have many virtual domains, each with its own access and error logfile. I'm using CustomLog, not TransferLog. Apache is chrooted. Adding every logfile to /etc/newsyslog.conf is one way, but hard to maintain. Is Apache's own rotatelogs program the way to go? Mackan
Re: Too many open files in system, /bsd: file: table is full
On 27 aug 2006, at 17.38, Stuart Henderson wrote: On 2006/08/27 17:09, Mackan wrote: I did "sysctl -w kern.maxfiles=1" to increase the value from 1772 to 1, and put the same value in /etc/sysctl.conf Keep an eye on kern.nfiles, if it's higher than expected ... It varies, mostly between 400-700. I guess it hit the limit at some peak. And is 1 a too high value? Depends, again. I'll go with 1. Just curious whether something bad could happen with a limit set too high. Thanks, Mackan
Too many open files in system, /bsd: file: table is full
Hi! I got a few "Postmaster notify" from Sendmail, telling me there was "mail.local: /var/mail/many: Too many open files in system" and also a few with "mail: can't load library 'libc.so.39.0'". /var/log/messages told me: Aug 27 16:00:01 blowfish /bsd: file: table is full Aug 27 16:00:01 blowfish last message repeated 19 times Aug 27 16:02:11 blowfish last message repeated 4 times Aug 27 16:07:14 blowfish last message repeated 128 times So, I made some googling, and found a few similar cases, and someone suggested changing the kern.maxfiles. I did "sysctl -w kern.maxfiles=1" to increase the value from 1772 to 1, and put the same value in /etc/sysctl.conf This is a GENERIC 3.9 install, on an i386 P4 1GB ram machine, running Sendmail, Apache, MySQL, Samba. I have a few scripts in crontab collecting SNMP data from my router to MRTG, besides that this is a low traffic server, with few users and a few websites. I think 1772 files seems to be too low as default. Is it? And is 1 a too high value? Mackan
Need help with dns/pf/tcpdump
Hi! Need help to understand dns/pf/tcpdump. See below. ns.foo.bar is a dns slave that makes AXFR zone transfer from my server (mybox). Why is the traffic blocked on the first lines? What kind of traffic is that? Perhaps I don't understand DNS fully, but I thought zone transfers were made using TCP only, and ordinary queries UDP. Here's the relevant part of my pf config: $ext_if = "bge0" rule 18) pass in log quick on $ext_if inet proto tcp from to any port = domain rule 19) pass in log quick on $ext_if inet proto udp from to any port = domain rule 21) block return in log on $ext_if all # tcpdump -a -e -o -ttt -i pflog0 Jul 24 05:34:15.090025 rule 21/(match) block in on bge0: ns.foo.bar. 55009 > mybox.domain: [|domain] (DF) Jul 24 05:34:19.089201 rule 21/(match) block in on bge0: ns.foo.bar. 55009 > mybox.domain: [|domain] (DF) Jul 24 05:34:27.090143 rule 21/(match) block in on bge0: ns.foo.bar. 55009 > mybox.domain: [|domain] (DF) Jul 24 05:34:43.107922 rule 18/(match) pass in on bge0: ns.foo.bar. 43646 > mybox.domain: [|tcp] (DF) Jul 24 05:34:46.101496 rule 18/(match) pass in on bge0: ns.foo.bar. 43646 > mybox.domain: [|tcp] (DF) Jul 24 05:34:52.101598 rule 18/(match) pass in on bge0: ns.foo.bar. 43646 > mybox.domain: [|tcp] (DF) Jul 24 05:35:04.102286 rule 18/(match) pass in on bge0: ns.foo.bar. 43646 > mybox.domain: [|tcp] (DF) Jul 24 05:35:28.102625 rule 18/(match) pass in on bge0: ns.foo.bar. 43646 > mybox.domain: [|tcp] (DF) No transfers were actually made this time, I guess the slaves just checking in and compares serial number of zones. But transfers do work according to logfiles from named. Anybody? :-) Mackan
Re: Make pf reload ruleset whenever a new file appears/changes
Martin Schrvder wrote: > 2006/7/18, Mackan <[EMAIL PROTECTED]>: >> 4) same php script generates a new ruleset for pf >> 5) pf detect changes and reload new ruleset >> >> Step 1 - 4 is already done. I need help with step 5. > > You know pfctl(8)? Yes. But how do I make apache/php execute the pfctl program or signal to pfctl ro reload? Apache is chroot and run by www, and pfctl lives outside chroot and must be run as root. I can't see your point - perhaps I'm missing something here. Mackan
Make pf reload ruleset whenever a new file appears/changes
Hi list! (Warning: poor english ahead!) This is what I need to do: 1) a user authenticates and enters my website 2) the user enters one IP-address into a form and submit it. 3) php script receive this address and save it in a file or database. 4) same php script generates a new ruleset for pf 5) pf detect changes and reload new ruleset Step 1 - 4 is already done. I need help with step 5. Apache is chroot. I could have a cron job search for a new file every five minutes or so, but that feels like an ugly solution. Is there any better way to signal pf to reload the new ruleset whenever a user submits a new IP-adress? Odd question perhaps... :-) Mackan
Re: UTF-8 text editor
On 10 jul 2006, at 21.54, Will H. Backman wrote: Mackan wrote: I just "upgraded" my server OS from Debian/Linux to OpenBSD. But it seems that in the case of Unicode-aware applications I made a big "downgrade". I really want I simple editor with unicode, for myself and my users. Mackan What was lacking from Debian/Linux that made you decide to switch to OpenBSD? Nothing at all. Debian/Linux is a very functional and feature-rich OS. But I like OpenBSD because of its clean design, and being a more UNIX'ish OS. I decided to use it for my new server. Everything currently running on the old Debian box is now running on the new OpenBSD box. And If I could find a working utf-8 console editor everything would be perfect! Perhaps "upgrade" was the wrong word to use. Mackan
Re: UTF-8 text editor
On 10 jul 2006, at 20.43, Spruell, Darren-Perot wrote: From: [EMAIL PROTECTED] Is there any UTF-8-aware text editor (for terminal use) available for OpenBSD? Vi(m) and similar is out of question for me, I never learned those. As ubiquitous as vi is on Unix, it seems a shallow reason. Really, it takes all of 15 minutes to pick up what you need for vi/ vim. Install a copy somewhere and spend a few minutes on vimtutor and you should find it pretty straightforward. DS You are probably right about that. We'll see. I just "upgraded" my server OS from Debian/Linux to OpenBSD. But it seems that in the case of Unicode-aware applications I made a big "downgrade". I really want I simple editor with unicode, for myself and my users. Mackan
UTF-8 text editor
Hi list! Is there any UTF-8-aware text editor (for terminal use) available for OpenBSD? Vi(m) and similar is out of question for me, I never learned those. I tried to compile latest nano from CVS, which support UTF-8, but with no luck. I get configure errors saying that my curses don't support unicode. Using 3.9/i386 with GENERIC. Suggestions anyone? Thanks, Mackan