eZ Community » Forums » General » Stuck cronjob process??
expandshrink

Stuck cronjob process??

Stuck cronjob process??

Thursday 23 August 2012 7:44:28 pm - 4 replies

Hi. Im having a cronjob process that takes almost 95% of one of the processors... It has been on for hours. and taking load up...

/usr/local/bin/php runcronjobs.php -q

I have killed it before, but I dont understan what is keeping it stuck this is new, never happened before, how may I know?

I have deleted caches, and checked a lot of things. Could it be php times? I have them in 240 sec.

FCGI maybe or what?

I have EZ 4.4 not sure if there are any changes with current PHP and MySQL versions (5.3.13 and 5.5)

or what could be a reason?.

MySQL server is almost idle.

Help here. thanks.

EDIT: More info: the main cronjob task is ran everyday at 4 am right now its 13:00 hrs here. So its been around 9 hours running. Just killed it... nothing happened. Somehow it is not ending its work on time. Load came down around 2 points.

Modified on Thursday 23 August 2012 8:05:04 pm by Luis D Garcia

Thursday 23 August 2012 10:21:49 pm

Take off the -q that might help...

Otherwise if it's on a linux system and you have command line access, the command strace is your friend.

Friday 24 August 2012 2:45:48 pm

Hi, Steven.

Yes I have command line access. How do I run strace?

If I remove the -q (quiet) does that mean that it is going to log something somewhere?

Today the problem is still there today...

Friday 24 August 2012 3:05:30 pm

Steven,

Using a trace from WHM in the process I found something that is cycling (but its not, I have found some differences between cycles) I removed database name and user name... I really dont understand a thing here:

munmap(0x2afc64c9f000, 4096)            = 0poll([{fd=4, events=POLLIN|POLLPRI}], 1, 0) = 0 (Timeout)write(4, "\306\0\0\0\3SELECT id, parent, lang_mas"..., 202) = 202read(4, "\1\0\0\1\5A\0\0\2\3def\rdatabase\rezur"..., 16384) = 528poll([{fd=4, events=POLLIN|POLLPRI}], 1, 0) = 0 (Timeout)write(4, "\375\10\0\0\3SELECT DISTINCT\n           "..., 2305) = 2305read(4, "\1\0\0\1!_\0\0\2\3def\rdatabase\17ezco"..., 16384) = 2736read(4, "rt_order\f?\0\v\0\0\0\3\0\0\0\0\0m\0\0\36\3def\rqu"..., 16384) = 534poll([{fd=4, events=POLLIN|POLLPRI}], 1, 0) = 0 (Timeout)write(4, "\325\0\0\0\3SELECT contentobject_state_"..., 217) = 217read(4, "\1\0\0\1\2q\0\0\2\3def\rdatabase\21ezco"..., 16384) = 227poll([{fd=4, events=POLLIN|POLLPRI}], 1, 0) = 0 (Timeout)write(4, "J\1\0\0\3SELECT l.contentobject_stat"..., 334) = 334read(4, "\1\0\0\1\3a\0\0\2\3def\rdatabase\1l\21ez"..., 16384) = 317poll([{fd=4, events=POLLIN|POLLPRI}], 1, 0) = 0 (Timeout)write(4, "\306\0\0\0\3SELECT id, parent, lang_mas"..., 202) = 202read(4, "\1\0\0\1\5A\0\0\2\3def\rdatabase\rezur"..., 16384) = 539poll([{fd=4, events=POLLIN|POLLPRI}], 1, 0) = 0 (Timeout)write(4, "9\0\0\0\3SELECT remote_id FROM ezcon"..., 61) = 61read(4, "\1\0\0\1\1S\0\0\2\3def\rdatabase\17ezco"..., 16384) = 147poll([{fd=4, events=POLLIN|POLLPRI}], 1, 0) = 0 (Timeout)write(4, "\376\5\0\0\3SELECT ezcontentobject.*,\n "..., 1538) = 1538read(4, "\1\0\0\1!_\0\0\2\3def\rdatabase\17ezco"..., 16384) = 2736read(4, "rt_order\f?\0\v\0\0\0\3\0\0\0\0\0m\0\0\36\3def\rqu"..., 16384) = 814access("var/ezflow_site/cache/template/compiled/generate-b7deefd8bdbef308cbc829a19c31bf6a.php", F_OK) = 0getcwd("/home/user/public_html/"..., 4096) = 28lstat("/home/user/public_html/./lib/ezc/var/ezflow_site/cache/template/compiled/generate-b7deefd8bdbef308cbc829a19c31bf6a.php", 0x7fffc8842f30) = -1 ENOENT (No such file or directory)getcwd("/home/user/public_html/v2"..., 4096) = 28open("/home/user/public_html/var/ezflow_site/cache/template/compiled/generate-b7deefd8bdbef308cbc829a19c31bf6a.php", O_RDONLY) = 8fstat(8, {st_mode=S_IFREG|0666, st_size=235548, ...}) = 0lseek(8, 0, SEEK_SET)                   = 0fcntl(8, F_GETFL)                       = 0x8000 (flags O_RDONLY|O_LARGEFILE)fstat(8, {st_mode=S_IFREG|0666, st_size=235548, ...}) = 0mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2afc64c9f000lseek(8, 0, SEEK_CUR)                   = 0open("/home/user/public_html/var/ezflow_site/cache/template/compiled/generate-b7deefd8bdbef308cbc829a19c31bf6a.php", O_RDONLY) = 9fstat(9, {st_mode=S_IFREG|0666, st_size=235548, ...}) = 0mmap(NULL, 235548, PROT_READ, MAP_SHARED, 9, 0) = 0x2afc64ca0000munmap(0x2afc64ca0000, 235548)          = 0close(9)                                = 0

Friday 24 August 2012 3:54:56 pm

The -q is for quiet - and cron should be generating a mail with the output of the command... but I think it gets written when the process finishes and since this is infinite you won't ever get that mail...

strace is a command... but it does pretty much what you displayed here so that was the point...

I'm seeing it trying to access a lot of cache, so I'm guessing it's maybe rebuilding the cache do you have the eztc.php script in your cronjobs?

It's possible that if you have a hanging bracket in one of your templates some thing could be doing something weird when the template gets compiled.

Otherwise, do you have a cluster setup/static setup?

I think this is the once a day cron - so if you run that command from the command line without the -q from your ezroot you can maybe get a clue where it's getting stuck (beware that people will get notifications when they aren't expecting them if they have them set to send once a day - it might also have unintended consequences if you have a pending newsletter or something like that).  If you can reproduce the problem on the command line, then you can try and comment out your Scripts[] in cronjob.ini.append.php one by one until you figure out which script is hanging.

If it's not the once a day cron that hangs then you can try the infrequent/frequent etc CronjobPart-<name> until you find the culprit.

If its an intermittent thing that only happens when the a group of criteria happen to be the same... it'll be hard to debug.

expandshrink

You must be logged in to post messages in this topic!

36 542 Users on board!

Forums menu

Proudly Developed with from