The problem is still there.

Can you please explain how to turn all vhosts to fastcgi?
Or just regenerating the config files after I switched to worker is enough?, since mod_php will not work anymore with worker.
And what about a2dismod mod_php?

Actually i don't see mod_php in my modules.
@joximu: how can you see in my previous posts, that some vhosts are using mod_php?

Loaded Modules:
core_module (static)
log_config_module (static)
logio_module (static)
mpm_worker_module (static)
http_module (static)
so_module (static)
actions_module (shared)
alias_module (shared)
auth_basic_module (shared)
authn_file_module (shared)
authz_default_module (shared)
authz_groupfile_module (shared)
authz_host_module (shared)
authz_user_module (shared)
autoindex_module (shared)
cband_module (shared)
cgid_module (shared)
dav_module (shared)
dav_fs_module (shared)
dav_svn_module (shared)
authz_svn_module (shared)
dir_module (shared)
env_module (shared)
fastcgi_module (shared)
info_module (shared)
mime_module (shared)
security2_module (shared)
python_module (shared)
negotiation_module (shared)
proxy_module (shared)
proxy_http_module (shared)
rewrite_module (shared)
setenvif_module (shared)
status_module (shared)
suexec_module (shared)
unique_id_module (shared)
watch_module (shared)

Sorry

no, it seems ok with the php thing.

About the midnight thing:
did you already switch the backup format to gzip instead of bzip2 (see /etc/ispcp/ispcp.conf)

and then, I'd suggest to move the backup to something like 3 o clock:

eg:
[ccode]
# Backup
0 3 * * * root /var/www/ispcp/engine/backup/ispcp-backup-all yes &>/var/log/ispcp/ispcp-backup-all-mngr.log
30 2 * * * root /var/www/ispcp/engine/backup/ispcp-backup-ispcp noreport &>/var/log/ispcp/ispcp-backup-ispcp-mngr.log
[/code]

bzip2 takes alot of cpu - you can search for the bzip2/gzip discussion in the forum (i think it was somewhere in january).

/J

as far as i see - its nothing wrong with the memory consuming - since this seems to be a graph out of munin - so far i know it just means that the mem is beeing used for cache - and thats normal....
Only the green color is mem your applications are really using - so no need to worry about that. Its normal linux / Unix behaviour to tell you its caching things in mem if there's enough available ... (instead of win32 platform) .

Well when I don't reboot it for a week, it's using some swap cache and the ram has only a few mb's left.
I don't think this is normal.

Thanx joximu for the cron explaination.
I'm gonna try it right away.

I just switched backups from bzip2 to gzip in the ispcp.conf

Also in /cron.d/ispcp I changed:


to


Lets see what happens now at 0:00, or hopefully at 3:00

hi prale

the memory "consumption" is normal in linux if it takes the whole you have - theres no need to "not use" any Byte if it's available? Why are you paying for 512MB of ram and then only use 400MB - only for you to be able to say: well, I still have enough RAM...

Unix takes the whole RAM - if needed. Windows (at least Vista I think) does this too but does not show this in the same way, "free" means "free or used for cache" in Redmond...

/J

yeah maybe you should make sure you've a bug-free php version ...
the debian etch / stable version has a bug which causes that child php - processes won't get killed if they've reached their lifetime.

Regarding your memory use : even your "top" output shows total normal values ...
500m used, ~340m cached ... and so on thats normal.

it's using 256.13 MB RAM now at 0:05
so let's see what happens at 2:30 and 3:00

Hopefully it will be more - otherwise the software does not use all of the hardware that is available... :-)