You are not logged in.
- Topics: Active | Unanswered
#1 2014-03-01 04:24:57
- yjcong
- Member
- Registered: 2009-07-06
- Posts: 59
[SOLVED]"at" does not work anymore
hi,
I have forgot what I did. "at" command has stopped working since yesterday, and I totally have no clue what I should do to fix it. There are new things I have found:
1)Everytime, after turning on the power and loginning, I notice a high cpu usage(60%-80%) by "atd"
2) "at -c" returns nothing, and these jobs have never been done
3)
warning: commands will be executed using /bin/sh
job 1 at Sat Mar 1 12:14:00 2014
warning: commands will be executed using /bin/sh
Cannot create atjob file /var/spool/atd/a0000101626c3e: File exists4)Those are job which I typed from different time. But all of them have the same ID
1 Sat Mar 1 12:05:00 2014 a yjc
1 Sat Mar 1 12:14:00 2014 = yjc
1 Sat Mar 1 12:14:00 2014 a yjc
1 Sat Mar 1 12:05:00 2014 = yjc
1 Sat Mar 1 12:12:00 2014 = yjc
1 Sat Mar 1 12:12:00 2014 a yjcThanks for the help.
Last edited by yjcong (2014-03-01 09:36:01)
Offline
#2 2014-03-01 06:41:46
- DaveCode
- Member
- Registered: 2008-08-15
- Posts: 103
Re: [SOLVED]"at" does not work anymore
The lovely at package saved me twice now, getting systemd to reboot a PC and launching wineserver decently on user X login.
All I can do is guess but high CPU smells like disk I/O with errors. My thought is spool thrashing or corrupt filesys. BACKUP esp. if you have lots of e-mail in spool. Then for a bypass test run at from tmpfs. What you do is:
/etc/fstab
tmpfs /tmp tmpfs rw,nosuid,exec,noatime,size=800M,nr_inodes=80k 0 0
tmpfs /run tmpfs defaults,noatime,size=40M 0 0
tmpfs /media tmpfs defaults,noatime,size=10M 0 0
tmpfs /var/log tmpfs defaults,noatime,size=50M 0 0
tmpfs /var/cache tmpfs defaults,noatime,size=150M 0 0
tmpfs /var/spool tmpfs defaults,noatime,size=40M 0 0
tmpfs /var/tmp tmpfs defaults,noatime,size=10M 0 0Edit to taste, but keep at least the /var/spool line for at.
Then do
/etc/tmpfiles.d/spool.conf
d /var/spool/mail 1777 root root 1d
d /var/spool/atd 1770 daemon daemon 1d
f /var/spool/atd/.SEQ 0600 daemon daemon -
d /var/spool/incron 0755 incron incron 1dDelete incron if you don't have it.
Reboot and see what happens. If at works, I think it's safe to conclude "disk problems."
Offline
#3 2014-03-01 07:41:20
- yjcong
- Member
- Registered: 2009-07-06
- Posts: 59
Re: [SOLVED]"at" does not work anymore
The lovely at package saved me twice now, getting systemd to reboot a PC and launching wineserver decently on user X login.
All I can do is guess but high CPU smells like disk I/O with errors. My thought is spool thrashing or corrupt filesys. BACKUP esp. if you have lots of e-mail in spool. Then for a bypass test run at from tmpfs. What you do is:
Thank you.
But I am a little bit confused.
there is nothing in /etc/tmpfiles.d/
/etc/fstab
UUID=fb94e9da-143a-4682-816f-259eb1daf22f / ext4 defaults,noatime 0 1
UUID=71fa21e7-3f42-46c0-b7a5-08fc8c8da4f6 /boot ext2 defaults,noatime 0 0
UUID=c1953a96-4cd8-429f-ace7-19daa9522be3 none swap defaults 0 0
UUID=eb6172b6-baf8-4b6b-877e-a780e9c0a12f /home reiserfs defaults,noatime 0 df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda1 9.7G 9.2G 0 100% /
dev 751M 0 751M 0% /dev
run 753M 712K 752M 1% /run
tmpfs 753M 12K 753M 1% /dev/shm
tmpfs 753M 0 753M 0% /sys/fs/cgroup
tmpfs 753M 468K 752M 1% /tmp
/dev/sda4 304M 34M 255M 12% /boot
/dev/sda5 19G 17G 1.8G 91% /home/var/spool# du -s -BM *
1M anacron
1M atd
1M cron
1M mailand I do not use mail to receive nor send emails.
I am not sure, in my case, spool directly connects with "at"
Last edited by yjcong (2014-03-01 07:51:01)
Offline
#4 2014-03-01 09:35:09
- yjcong
- Member
- Registered: 2009-07-06
- Posts: 59
Re: [SOLVED]"at" does not work anymore
Thanks DaveCode.
I figure out the reason. It is not spool, but as you said " smells disk I/O with errors". I used bleachbit to clean /, and pressed Ctrl-C before it finished. so there is a huge folder in /root which eats up all of free space.
Offline
#5 2014-03-01 23:53:45
- DaveCode
- Member
- Registered: 2008-08-15
- Posts: 103
Re: [SOLVED]"at" does not work anymore
Glad things worked out. Disk-full scenarios can glitch many things.
I am a little bit confused
By default nothing lives in /etc/tmpfiles.d but you add new confs there.
I am not sure, in my case, spool directly connects with "at"
Then inspect the bottom of the at package file list. The daemon needs its stuff extant to start properly. Anything in tmpfs disappears at shutdown so that stuff needs to be remade at boot. Here's how:
man tmpfiles.dYour root disk is ext4 journaled. I recommend a full backup then reformatting without ext4 journal. From a boot CD you can use something like
mkfs.ext4 -v -t ext4 -i 8192 -O ^has_journal -L MyRootDiskName -U fb94e9da-143a-4682-816f-259eb1daf22f /dev/disk/by-uuid/fb94e9da-143a-4682-816f-259eb1daf22fTweak other ext4 defaults inside /etc/mke2fs.conf as you may please. (Meaning the one on the boot CD.)
If you're having disk issues then putting some items in tmpfs as shown may help alleviate them. You will also get faster system performance.
Last edited by DaveCode (2014-03-01 23:55:48)
Offline
#6 2014-03-04 11:31:39
- yjcong
- Member
- Registered: 2009-07-06
- Posts: 59
Re: [SOLVED]"at" does not work anymore
Your root disk is ext4 journaled. I recommend a full backup then reformatting without ext4 journal. From a boot CD you can use something like
mkfs.ext4 -v -t ext4 -i 8192 -O ^has_journal -L MyRootDiskName -U fb94e9da-143a-4682-816f-259eb1daf22f /dev/disk/by-uuid/fb94e9da-143a-4682-816f-259eb1daf22fTweak other ext4 defaults inside /etc/mke2fs.conf as you may please. (Meaning the one on the boot CD.)
If you're having disk issues then putting some items in tmpfs as shown may help alleviate them. You will also get faster system performance.
Thanks for your recommendation. I will put it in my TODO list 
Offline
#7 2014-03-04 11:45:42
- falconindy
- Developer
- From: New York, USA
- Registered: 2009-10-22
- Posts: 4,111
- Website
Offline