Page 1 of 1
crontab
Posted: Tue Mar 06, 2018 4:43 am
by nhdesign
Copied my crontab from bolt to sh yesterday.
31 emails from (cron daemon) last night.
Not even a touch command worked
touch: cannot touch ???/nfs/ftp/pub/users/nhdesign/outgoing/33660-52-foot-shift-crank.jpg???: No such file or directory
Crontab is gone.
Presume someone deleted it.
When can I reload the crontab?
Thanks,
Vic
Re: crontab
Posted: Tue Mar 06, 2018 10:48 am
by scott
nhdesign wrote:Copied my crontab from bolt to sh yesterday.
31 emails from (cron daemon) last night.
Not even a touch command worked
touch: cannot touch ???/nfs/ftp/pub/users/nhdesign/outgoing/33660-52-foot-shift-crank.jpg???: No such file or directory
Crontab is gone.
Presume someone deleted it.
When can I reload the crontab?
Thanks,
Vic
Hi Vic,
I'm looking into this now. Last night, the server had another problem with resources. Two things were going on: first, your jobs in your crontab were being denied by SELinux (security enhanced Linux), and second, certain mounts for the resource management system known as "cgroups" had somehow disappeared from the system. One of our system admins rebooted the system last night, and it very shortly ended up in the same situation. He moved two crontabs aside (including yours), then tried again, and the system worked as expected.
So it seems there is something amiss with the way the system sets up crontab environments, which with busy crontabs can cause the system to lose its mind. So I'm bird-dogging that now.
-Scott
Re: crontab
Posted: Tue Mar 06, 2018 7:44 pm
by scott
scott wrote:nhdesign wrote:Copied my crontab from bolt to sh yesterday.
31 emails from (cron daemon) last night.
Not even a touch command worked
touch: cannot touch ???/nfs/ftp/pub/users/nhdesign/outgoing/33660-52-foot-shift-crank.jpg???: No such file or directory
Crontab is gone.
Presume someone deleted it.
When can I reload the crontab?
Thanks,
Vic
Hi Vic,
I'm looking into this now. Last night, the server had another problem with resources. Two things were going on: first, your jobs in your crontab were being denied by SELinux (security enhanced Linux), and second, certain mounts for the resource management system known as "cgroups" had somehow disappeared from the system. One of our system admins rebooted the system last night, and it very shortly ended up in the same situation. He moved two crontabs aside (including yours), then tried again, and the system worked as expected.
So it seems there is something amiss with the way the system sets up crontab environments, which with busy crontabs can cause the system to lose its mind. So I'm bird-dogging that now.
-Scott
Your crontab is back in place.
This new change to the system that I worked on all day should mitigate a lot of the problems. There's still one corner case I need to fix with the automated unmounting, but I think we should last overnight with this setup (fingers crossed, knock on wood).
-Scott
Re: crontab
Posted: Wed Mar 07, 2018 11:27 pm
by lr
Copied my crontab over today. Completely painless. Boring. Worked perfectly.
Re: crontab
Posted: Thu Mar 08, 2018 5:31 am
by nhdesign
I deleted mine and put it back on bolt until the dust settles. Oddly I have had an issue with bolt. Is there a chance you changed something on bolt? Both nights, two nights, fortunately I run backups. Maybe it is related to the missing files thing?
--
Vic
Re: crontab
Posted: Thu Mar 08, 2018 10:17 am
by patty1
I've been running crontab on both machines. I only use it to grab entries from a calendar file in my home directory. I put a three-times-a-week entry in my calendar just to see how things are going. Two nights ago, sh.sonic.net returned a "file not found" error, but last night, both machines worked fine. Since I only run one cron job a day, I'm not as apt to catch problems as heavier users are.
Re: crontab
Posted: Thu Mar 08, 2018 11:19 am
by goetsch
I've got a job that runs hourly; last error seen yesterday (Wednesday) in the 12:00pm hour, so we're coming up on 23 error-free hours...
Re: crontab
Posted: Thu Mar 08, 2018 12:05 pm
by patty1
Thanks, goetsch, your hourly cron job is a much better indicator of sh's health than my daily one.
Re: crontab
Posted: Thu Mar 08, 2018 12:57 pm
by goetsch
Another cron job, another fail...
Every day at 12:31pm I copy the previous day's webserver access log for a Sonic-hosted domain to a folder in my home directory. Today's error was for a file not found. This server log should live at
Code: Select all
/var/log/httpd/[user name]/[hosted domain name]/access_log
When I ssh into sh.sonic.net and start walking down the directory tree, I find that there is no
httpd directory under
/var/log
Code: Select all
$ ls /var
account cache cvs empty gopher lib lock mail opt run tmp
adm crash db games kerberos local log nis preserve spool yp
Code: Select all
$ ls /var/log
anaconda btmp denyhosts-20180203 mail messages-20180304 secure-20180225 wtmp
audit btmp-20180301 denyhosts-20180220 maillog munin-node secure-20180304 wtmp-20180304
boot.log chrony denyhosts-20180301 maillog-20180211 none shell yum.log
boot.log-20180302 cron denyhosts-20180307 maillog-20180218 ntpstats spooler yum.log-20180101
boot.log-20180303 cron-20180211 dmesg maillog-20180225 ppp spooler-20180211
boot.log-20180304 cron-20180218 dmesg.old maillog-20180304 rhsm spooler-20180218
boot.log-20180305 cron-20180225 grubby messages sa spooler-20180225
boot.log-20180306 cron-20180304 grubby_prune_debug messages-20180211 secure spooler-20180304
boot.log-20180307 custweb iptraf-ng messages-20180218 secure-20180211 tallylog
boot.log-20180308 denyhosts lastlog messages-20180225 secure-20180218 tuned
Code: Select all
$ ls /var/log/httpd
ls: cannot access /var/log/httpd: No such file or directory
I'm running the same cron jobs on bolt, so the copy is taking place (I'm getting my file), but no thanks to sh.sonic.net.
Re: crontab
Posted: Thu Mar 08, 2018 1:28 pm
by scott
goetsch wrote:Another cron job, another fail...
Every day at 12:31pm I copy the previous day's webserver access log for a Sonic-hosted domain to a folder in my home directory. Today's error was for a file not found. This server log should live at
Code: Select all
/var/log/httpd/[user name]/[hosted domain name]/access_log
When I ssh into sh.sonic.net and start walking down the directory tree, I find that there is no
httpd directory under
/var/log
Code: Select all
$ ls /var
account cache cvs empty gopher lib lock mail opt run tmp
adm crash db games kerberos local log nis preserve spool yp
Code: Select all
$ ls /var/log
anaconda btmp denyhosts-20180203 mail messages-20180304 secure-20180225 wtmp
audit btmp-20180301 denyhosts-20180220 maillog munin-node secure-20180304 wtmp-20180304
boot.log chrony denyhosts-20180301 maillog-20180211 none shell yum.log
boot.log-20180302 cron denyhosts-20180307 maillog-20180218 ntpstats spooler yum.log-20180101
boot.log-20180303 cron-20180211 dmesg maillog-20180225 ppp spooler-20180211
boot.log-20180304 cron-20180218 dmesg.old maillog-20180304 rhsm spooler-20180218
boot.log-20180305 cron-20180225 grubby messages sa spooler-20180225
boot.log-20180306 cron-20180304 grubby_prune_debug messages-20180211 secure spooler-20180304
boot.log-20180307 custweb iptraf-ng messages-20180218 secure-20180211 tallylog
boot.log-20180308 denyhosts lastlog messages-20180225 secure-20180218 tuned
Code: Select all
$ ls /var/log/httpd
ls: cannot access /var/log/httpd: No such file or directory
I'm running the same cron jobs on bolt, so the copy is taking place (I'm getting my file), but no thanks to sh.sonic.net.
Take a look in /logs/by_user -- you'll find your domain's directories in there.
-Scott