This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
Pages: 1 [2]
16
Installation and Upgrades / Re: ebox memory footprint
« on: October 08, 2009, 02:54:44 pm »Just a thought, the output of top that you posted doesn't seem to be ordered by memory use. You have to press "M" when top is running to do that (the default is to order by cpu use and lots of services site idle most of the time so you don't see them on that list)
here is top from one of my servers:Code: [Select]top - 22:28:16 up 50 days, 11:25, 2 users, load average: 0.00, 0.00, 0.00
Tasks: 124 total, 2 running, 122 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 1017960k total, 980308k used, 37652k free, 122600k buffers
Swap: 2980016k total, 908k used, 2979108k free, 316976k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
8482 clamav 20 0 86456 73m 980 S 0.0 7.4 2:48.57 clamd
8542 amavis 20 0 66852 56m 5816 S 0.0 5.7 0:05.67 amavisd-new
30105 amavis 20 0 69156 56m 2660 S 0.0 5.6 0:01.72 amavisd-new
30788 amavis 20 0 68464 55m 2560 S 0.0 5.6 0:00.79 amavisd-new
7465 ebox 20 0 51488 44m 4748 S 0.3 4.5 106:46.78 ebox-loggerd
13850 ebox 20 0 53832 43m 5652 S 0.0 4.4 0:02.02 apache2
2032 www-data 20 0 46372 28m 5972 S 0.0 2.9 3:29.19 apache2
8507 root 20 0 31584 28m 2572 S 0.0 2.9 0:52.57 spamd
8567 root 20 0 31584 26m 640 S 0.0 2.7 0:00.27 spamd
8568 root 20 0 31584 26m 548 S 0.0 2.7 0:00.38 spamd
1620 www-data 20 0 42356 24m 6012 S 0.0 2.5 8:18.92 apache2
14401 www-data 20 0 41752 23m 5624 S 0.0 2.4 1:27.62 apache2
14429 www-data 20 0 40776 23m 5996 S 0.0 2.3 3:21.00 apache2
8512 ebox 20 0 26048 22m 3364 S 0.0 2.3 0:03.76 learnspamd
7695 postgres 20 0 40812 22m 20m S 0.0 2.2 3:38.85 postgres
4887 postgres 20 0 40144 21m 21m S 0.0 2.2 2:32.74 postgres
27700 www-data 20 0 39244 20m 5536 S 0.0 2.1 0:26.95 apache2
As you can see clamAV is the worst offender along with anti spam stuff, apache and postgres (but then I do use egroupware). Given it's been up for 50 days I don't think the memory usage is too bad, of course I'm not acting as a proxy for 200 clients though.
If you want a real memory hog have a look at a server2008 SBS install. It will boot with 4GB but you seem to need more like 8GB if you don't want it to keep crashing... and that's with 5 users
ok, but i dont use egroupware, mail, antispam, antivirus, database (postgres) so i didnt install that modules, but still the memory used is very high. I dont want to use window$, i can compare with other linux distros like ipcop that the only issue is that not support multiwan, and giving all same services that ebox use only 1/8 of the memory that ebox use.
I think there are 3 things that cause the mem problem:
1- the soft that ebox use to each service (ej. ebox for dns use bind and ipcop use dnsmasq).
2- Use the ubuntu kernel without tunning (if somebody used ubuntu anytime knows what about iam taking, the ubuntu kernel use a lot of the ram to buffers and cache, i dont why).
3- Dependencies, ex. i dont select postgres in advance packages selection but it installed anyway, and the same thing do with other packages.
thanks and sorry about may english.
17
Installation and Upgrades / Re: multigateway + loadbalancer + wan failover
« on: October 07, 2009, 08:29:48 pm »
I want to know what are the limitation in multiwan because i have 2 modem-router for the isps and if i use them i get and static ip on each wan for the ebox, that works?
Something like this:
---isp1--modem1(doing nat)--10.0.0.1---ebox-eth1-----Multiwan1
----isp2--modem2(doing nat)--10.0.1.1--ebox-eth2-----Multiwan2
192.168.0.1-----ebox-eth0-----Lan
thanks and sorry about the english
Something like this:
---isp1--modem1(doing nat)--10.0.0.1---ebox-eth1-----Multiwan1
----isp2--modem2(doing nat)--10.0.1.1--ebox-eth2-----Multiwan2
192.168.0.1-----ebox-eth0-----Lan
thanks and sorry about the english
18
Installation and Upgrades / Re: multigateway + loadbalancer + wan failover
« on: October 06, 2009, 05:38:04 pm »QuoteSo i need static IPs on WAN conections, ok but there are no way to doit with dinamic IPs or in a future ebox will be ship this feature?
Maybe in the future we will support this, but not right now.QuoteBut i get a doubt, if gateway2 is set to only allow a protocol or other restriccion, anyway when gateway1 is down gateway2 take the control of all other traffic? because thats will be nice to me, i need that.
You are right, that's exactly what should happen in that case.
Regards,
J. A. Calvo
I have i question about of using dymamic ips on multiwan config, why we cant use dynamics ip? where are the limitation? because if we say that x trafic go through ethx is ready, why we need a static ip?
thanks
19
Installation and Upgrades / Re: ebox memory footprint
« on: October 06, 2009, 04:20:31 pm »It is not eating 500mb, you have to substract the buffered and cached memory that in fact they are also available. So your real used memory is:
451088-21480-191796=237812 (232Mb)
I don't know which software installs ipcop, but I think 232MB is not too much.
And looking at your top, apache is the main responsible of that memory consumption, so if you are going to use it as gateway I don't see any relation. I mean, maybe with 200 clients eBox will perform well if these clients are not using the ebox apache (they won't).
Hope this clarifies something.
Regards,
J. A. Calvo
Iam back, ok i know that buffers and cache is not real mem used but also is true that the kernel never free that mem when the phisical memory is running out (i know by experience), so start swapping , and also that the ipcop take only 50-60 mb (incluided buffers and cache with kernel 2.4.35 in stable release) and 60-70 (incluided buffers and cache with kernel 2.6.27 in unstable release).
Also if we take out apache the mem used is still huge.
thanks
bye
20
Installation and Upgrades / Re: ebox memory footprint
« on: October 02, 2009, 09:19:47 pm »Hi lucho,
First of all, we should know how many ebox modules do you have installed, so if you can please paste here the output of the following command: dpkg -l | grep ebox
It would be also nice if you give us the output of the top command to see the state of the running processes and the memory.
Regards,
J. A. Calvo
ok
Code: [Select]
dpkg -l | grep ebox
ii ebox 1.2.8-0ubuntu1~ppa1~hardy1 the eBox platform - Base framework
ii ebox-ca 1.2-0ubuntu1~ppa1~hardy1 eBox - Certificate Authority Manager for eBo
ii ebox-dhcp 1.2-0ubuntu1~ppa1~hardy1 eBox - DHCP server module
ii ebox-dns 1.2.2-0ubuntu1~ppa1~hardy1 eBox - DNS server
ii ebox-firewall 1.2.2-0ubuntu1~ppa1~hardy1 eBox - Firewall
ii ebox-l7-protocols 1.2-0ubuntu1~ppa1~hardy1 eBox - layer 7 protocols
ii ebox-monitor 1.2.2-0ubuntu1~ppa1~hardy1 eBox - Monitoring module
ii ebox-network 1.2.6-0ubuntu1~ppa1~hardy1 eBox - Network configuration module
ii ebox-ntp 1.2-0ubuntu1~ppa1~hardy1 eBox - NTP server
ii ebox-objects 1.2-0ubuntu1~ppa1~hardy1 eBox - Object management
ii ebox-openvpn 1.2.2-0ubuntu1~ppa1~hardy1 eBox - OpenVPN server module
ii ebox-services 1.2.1-0ubuntu1~ppa1~hardy1 eBox - Services management
ii ebox-software 1.2-0ubuntu1~ppa1~hardy2 eBox - Software management
ii ebox-squid 1.2.4-0ubuntu1~ppa1~hardy1 eBox - Proxy cache and content filter
ii ebox-trafficshaping 1.2-0ubuntu1~ppa1~hardy1 eBox - Traffic shaper for eBox
ii ebox-usersandgroups 1.2.2-0ubuntu1~ppa1~hardy1 eBox - User and Group management
ii esofttool 0.5.4 ebox software tool
ii libebox 1.2.1-0ubuntu1~ppa1~hardy1 eBox common library for server and client
ii samba-vscan 0.3.6cbeta5ebox1-2 Samba virus scanning VFS module
and Top:
Code: [Select]
Tasks: 75 total, 3 running, 72 sleeping, 0 stopped, 0 zombie
Cpu(s): 5.7%us, 0.3%sy, 0.0%ni, 93.7%id, 0.3%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 1018976k total, 451088k used, 567888k free, 21480k buffers
Swap: 1686784k total, 0k used, 1686784k free, 191796k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
6264 ebox 20 0 55036 43m 5392 S 5.0 4.4 0:04.03 apache2
4692 ebox 20 0 4240 2484 1796 S 0.3 0.2 0:02.20 gconfd-2
6333 root 20 0 63852 11m 2160 S 0.3 1.1 0:00.53 collectd
1 root 20 0 1808 772 548 S 0.0 0.1 0:01.36 init
2 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kthreadd
3 root RT -5 0 0 0 S 0.0 0.0 0:00.00 migration/0
4 root 15 -5 0 0 0 R 0.0 0.0 0:00.02 ksoftirqd/0
5 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/0
6 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 events/0
7 root 15 -5 0 0 0 S 0.0 0.0 0:00.01 khelper
41 root 15 -5 0 0 0 S 0.0 0.0 0:00.03 kblockd/0
44 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kacpid
45 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kacpi_notify
118 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kseriod
156 root 20 0 0 0 0 S 0.0 0.0 0:00.00 pdflush
157 root 20 0 0 0 0 S 0.0 0.0 0:00.02 pdflush
158 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kswapd0
202 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 aio/0
1368 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 ata/0
1374 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 ata_aux
1377 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 ksuspend_usbd
1384 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 khubd
1391 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 scsi_eh_0
1396 root 15 -5 0 0 0 S 0.0 0.0 0:00.01 scsi_eh_1
2103 root 15 -5 0 0 0 S 0.0 0.0 0:00.01 scsi_eh_2
2105 root 15 -5 0 0 0 S 0.0 0.0 0:00.01 scsi_eh_3
2364 root 15 -5 0 0 0 S 0.0 0.0 0:00.18 kjournald
This info is with the system just booted.
21
Installation and Upgrades / ebox memory footprint
« on: October 02, 2009, 08:47:32 pm »
HI, i want to use ebox only like a powerfull gateway, i has been using ipcop + plugins for 5 years, now i want to build an server (P4 2.4mhz, 1gb ram, 3 nics), with ipcop stable (1.4.21) the footprint was of 50 or 60 MB with no charge, and 250 or 260 with full charge (200 pcs), but the first that i see when install ebox is that the fresh install without charge eat 500 mb , iam testing this ebox server so i dont know how memory eat with full charge (200 pcs) but i start thinking that all the ram will be eaten in seconds and the server will swap all the time.
Why the huge diference of memory footprint between ipcop and ebox?
Anybody have in production a similar server with a huge charge?
thanks and sorry about my english
Why the huge diference of memory footprint between ipcop and ebox?
Anybody have in production a similar server with a huge charge?
thanks and sorry about my english
22
Installation and Upgrades / Re: multigateway + loadbalancer + wan failover
« on: October 02, 2009, 05:49:09 pm »Yes, if you have WAN failover enabled and one of the router fails, all the traffic will go for the other. And the recovery of the failed gateway will be also detected automatically.
And yes, I'm afraid that you can have problems using the multigateway features if the IPs of your routers are dynamic...
Regards,
J. A. Calvo
So i need static IPs on WAN conections, ok but there are no way to doit with dinamic IPs or in a future ebox will be ship this feature?
In the first post i wrote :
"for example, if i set that with gateway1 go http requests, and then gateway1 fail, all http request go to gateway2 until gateway1 was up again?"
You response:
"Yes, if you have WAN failover enabled and one of the router fails, all the traffic will go for the other. And the recovery of the failed gateway will be also detected automatically."
But i get a doubt, if gateway2 is set to only allow a protocol or other restriccion, anyway when gateway1 is down gateway2 take the control of all other traffic? because thats will be nice to me, i need that.
thanks and sorry again about my english.
23
Installation and Upgrades / multigateway + loadbalancer + wan failover
« on: October 01, 2009, 08:46:41 pm »
HI, i want to use ebox only like a powerfull gateway, i has been using ipcop + plugins for 5 years, now i want to build an server (P4 2.4mhz, 1gb ram, 3 nics), with 2 wan (dinamic ip), i need that the server do multigateway + loadbalance + wan failover, that works ok in ebox?
For example, if i set that with gateway1 go http requests, and then gateway1 fail, all http request go to gateway2 until gateway1 was up again?
There are any problem with dinamics ips on wans?
thanks and sorry about my english
For example, if i set that with gateway1 go http requests, and then gateway1 fail, all http request go to gateway2 until gateway1 was up again?
There are any problem with dinamics ips on wans?
thanks and sorry about my english
Pages: 1 [2]