Linux - GeneralThis Linux forum is for general Linux questions and discussion.
If it is Linux Related and doesn't seem to fit in any other forum then this is the place.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
I had a problem today with my sendmail...none of my clients were able to check their mail, including us here in the office. Anyways, I figure restarting the service would correct this problem. No. Once I restarted the service it said that it started except it stated:
"warning: /var/lock/subsys/????: no space left on device!, So begining to panic i rebooted the machine and upon reboot, all of my services began getting the same error, httpd, smb, etc...what the hell is going on, i cleared my tmp directory and that seemed to fix it, however it quickly began again. I getting to my last nerve now.
Did you check to see that all of the mount points are there? ie: what is in your /etc/fstab file? If it's a mail server there's a good chance your /var or /var/spool is on its own partition. Did you set the box up yourself or was the admin job handed off to you?
run a "df -i" on ure partition and itll tell u if ure out of inodes. unfortunately u cant just add inodes like water. u'll need to reformat that partition & specify another blocksize.
Originally posted by unSpawn run a "df -i" on ure partition and itll tell u if ure out of inodes. unfortunately u cant just add inodes like water. u'll need to reformat that partition & specify another blocksize.
That would indeed be a possible approach. I'd recommed to switch to ReiserFS while he's at it, saving tons of space with all those small files and speeding up things as well.
now this is crazy. I figure out that it is actually a program called interchange an e-commmerce server appending hundreds of logs in a crazy format, something like error~1.log.gz.1.gz.2.gz.3...., the directory itself is impossible to manipulate. you can't remove files or copy them, both commands with give you an error like "too many arguments..." it's kinda crazy, but I was able to successfully recover all the inodes by using midnight comm to just remove the entire directory but it went on forever and slowed the system down enourmously, so i stop it. the problem that i have now is i can't really do anything with the directory. Crazy stuff
now this is crazy. I figure out that it is actually a program called interchange an e-commmerce server appending hundreds of logs in a crazy format, something like error~1.log.gz.1.gz.2.gz.3...., the directory itself is impossible to manipulate. you can't remove files or copy them, both commands with give you an error like "too many arguments..." it's kinda crazy, but I was able to successfully recover all the inodes by using midnight comm to just remove the entire directory but it went on forever and slowed the system down enourmously, so i stop it. the problem that i have now is i can't really do anything with the directory. Crazy stuff
Interchange has an option to cleanup cached pages - i think it's in the sessions directory. Where Interchange lives. I've been wrestiling with it for about 2 weeks and the friday deadline and the current workload could create enough input to fuel an improbability drive through at least a couple of galaxy tours.
Interchange has a nagging bit at the end of the 'makecat' script about cleaning this up. grep it and be wiser than me.
Cheers
(edit)
And MOST DEFINATELY make use of reiserfs. Highly recommended for your situation.
I don't know very much about interchange and it was being setup by another admin in our office. I've been reading the docs for it but it isn't getting me anywhere.
Originally posted by elucid construct.error.log.1.gz.2.3.gz...
I don't know very much about interchange and it was being setup by another admin in our office. I've been reading the docs for it but it isn't getting me anywhere.
Sorry - i meant the cached _user_ sessions in the the catalog directories. they can get nasty if the cgi goes thunk. However this is not the problem you are facing.
If the directories are totally bloated and contain a lot of files you can expect to wait a while for the cleanup. Kill interchange until you can fix your space problem and find out what's going on in the catalogs that are generating all those error logs. (reading one or two of them might help)
That the thing, all the log files are empty in the directory giving me the problem. You open it in vi and poof...nothing. The files exist, just nothing in them. I had already killed the interchange daemon so it would stop appending log files. And since they are empty, it's not giving me much to go by...
Originally posted by elucid That the thing, all the log files are empty in the directory giving me the problem. You open it in vi and poof...nothing. The files exist, just nothing in them. I had already killed the interchange daemon so it would stop appending log files. And since they are empty, it's not giving me much to go by...
Right. But depending on how that partition was formatted you can chew up space with 'zero' size files as each file will take up at least [inode density] space. If it was formatted with the stadard setup it should be 4096 bytes, so even an empty file will still use 4K of space. I would 'rm -r <log_file_directory>' and wait until it completes. then re-create the directory with the proper permissions. That should at least clear up that problem.
Use reiserFS.
What tasks will the server be perorming? Looks like httpd and mail. It can bog down pretty quick with a few users. You might consider getting another couple (well 3 anyway) of SCSI drives so you can gain a little bit of performance by using different drives for some of the busier partiions. (Just my $.02)
At the very least I would do the following: (you should be able to do this while still keeping your root on the original drive.). Pleae keep in mind that this is the simplified version and that you should backup before you attmept anything like this.
1) get reiserfs working (at the kernel level)
2) Take some time to re-examine the tasks that server will be performing and decide where to put the 'busier' partitions. (/var/spool, /usr/local (or wherever your www pages are) and a partition for the interchange catalogs.) IMHO you should look at using a total of four drives. Partition and format the new partitions wih reiserfs.
3) Mount one of the partitions. Copy the desired directories to their new partition. I use this from within the folder I want to copy - 'find . -xdev | cpio -pm /<where_you_mounted_your_new_parttion>'. umount and mount the next new partition. Lather, rinse repeat.
4) change your fstab to mount the new partitions to the proper dirs.
It's a little more difficult to get your root fs to reiserfs as you'll have to do some extra shuffling. It's possible and not a bad idea.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.