Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In Register

Quick Links


Categories

In this Discussion

› › Would you buy an ultra LEB? - Page 3
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Would you buy an ultra LEB?

13»

Comments

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @rds100 said: @Damian 2.6.32-042stab049.6 is different (different memory accounting than the "standard" UBC based OpenVZ).

    Yea, it crashes just by restarting a CT way too many times in a row :)

    Francisco

  • dmmcintyre3dmmcintyre3 Member

    @ramnet said: Mostly it's the .rpm based distros that take up tons of memory.

    I have no trouble with yum and rpm on a 64mb VPS.

  • dmmcintyre3dmmcintyre3 Member

    @KuJoe: I can get CentOS 6 down to 4mb on OpenVZ.

    [root@vz ~]# ps aux
USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
root         1  0.0  1.0   2864  1312 ?        Ss   06:43   0:00 init
root     12199  0.0  0.3   2744   504 ?        Ss   09:26   0:00 /usr/sbin/dropbear
root     12203  0.1  0.9   2932  1212 ?        Rs   09:27   0:00 /usr/sbin/dropbear
root     12223  0.0  1.2   6244  1628 pts/0    Ss   09:27   0:00 -bash
root     12270  0.0  0.7   6052  1016 pts/0    R+   09:29   0:00 ps aux
[root@vz ~]# free -m
             total       used       free     shared    buffers     cached
Mem:           128          4        123          0          0          0
-/+ buffers/cache:          4        123
Swap:            0          0          0
  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @dmmcintyre3 said: @KuJoe: I can get CentOS 6 down to 4mb on OpenVZ.

    That's pretty sweet, .32 kernel?

    Francisco

  • dmmcintyre3dmmcintyre3 Member

    @Francisco said: .32 kernel?

    no, 2.6.18-308.el5.028stab099.3

  • yomeroyomero Member

    @dmmcintyre3 said: no, 2.6.18-308.el5.028stab099.3

    OMFG

    But no cron and syslog >_>

  • raindog308raindog308 Administrator, Veteran

    @dmmcintyre3 said: I can get CentOS 6 down to 4mb on OpenVZ.

    Do you think you might shave some more memory by using pdksh or dash instead of bash?

  • dmmcintyre3dmmcintyre3 Member

    Replaced bash with dash and added crond and syslog-ng, sitting at 6mb.

    # ps aux
USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
root         1  0.0  0.4   2864  1304 ?        Ss   10:16   0:00 init
root     15918  0.0  0.1   2744   516 ?        Ss   10:16   0:00 /usr/sbin/dropbear
root     15940  0.0  0.4   7036  1136 ?        Ss   10:16   0:00 crond
root     16183  0.1  0.3   2932  1028 ?        Ss   10:23   0:00 /usr/sbin/dropbear
root     16184  0.0  0.2   2064   556 pts/0    Ss   10:23   0:00 -dash
root     16205  0.0  0.1   3868   488 ?        S    10:23   0:00 supervising syslog-ng              
root     16206  0.0  0.6   4416  1724 ?        Ss   10:23   0:00 syslog-ng -p /var/run/syslog-ng.pid
root     16212  0.0  0.3   6052  1008 pts/0    R+   10:24   0:00 ps aux
# free -m
             total       used       free     shared    buffers     cached
Mem:           256          6        249          0          0          0
-/+ buffers/cache:          6        249
Swap:            0          0          0
  • yomeroyomero Member

    Dash is overkiller u_u
    Almost unusable if you are used to autocomplete weird filenames

  • dmmcintyre3dmmcintyre3 Member

    @yomero said: Dash is overkiller u_u

    Almost unusable if you are used to autocomplete weird filenames

    still 6mb with mksh

    # ps aux                                                                       
USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
root         1  0.0  0.4   2864  1296 ?        Ss   10:31   0:00 init
root      9320  0.0  0.1   2744   520 ?        Ss   10:31   0:00 /usr/sbin/dropbear
root      9327  0.0  0.4   7028  1120 ?        Ss   10:31   0:00 crond
root      9372  0.3  0.3   2932  1028 ?        Ss   10:32   0:00 /usr/sbin/dropbear
root      9373  0.0  0.2   2216   728 pts/0    Rs   10:32   0:00 -mksh
root      9400  0.0  0.1   3868   492 ?        S    10:32   0:00 supervising syslog-ng              
root      9401  0.0  0.6   4416  1732 ?        Ss   10:32   0:00 syslog-ng -p /var/run/syslog-ng.pid
root      9406  0.0  0.3   6052  1008 pts/0    R+   10:33   0:00 ps aux
# free -m                                                                                                                                            
             total       used       free     shared    buffers     cached
Mem:           256          6        249          0          0          0
-/+ buffers/cache:          6        249
Swap:            0          0          0
  • raindog308raindog308 Administrator, Veteran

    pdksh? I admit I've never benchmarked the memory usage of the shells, but I know that's what Hostigation uses for its 64/96 backup VPSes. Autocompletion works just fine with pdksh.

  • exussumexussum Member

    yeah i use pdksh quite a bit. nice shell

  • BoltersdriveerBoltersdriveer Member, LIR

    IPs are relatively ok-ish, but I don't foresee too many people getting 32mb plans.

  • quirkyquarkquirkyquark Member

    @Francisco said: Yea, it crashes just by restarting a CT

    From one of @Prometeus VPS:

    root@fire:~# uname -r
2.6.32-042stab053.5

    Maybe he can share his secrets for getting the very latest .32 OVZ kernels running stably? I'm guessing his nodes aren't fully loaded yet...

    Thing is, Fran, that the coming Ubuntu 12.04 LTS will require 2.6.24 or newer, because of minimum GLIBC requirements --- it won't work on 2.6.18, but works great on .32. I don't know what percentage of your clients use Ubuntu, but I imagine a lot of them will want to migrate to 12.04, especially those running 10.04 LTS....

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @quirkyquark said: From one of @Prometeus VPS:

    There isn't any 'magic', it just means his users aren't as brutal as they could be :) There's plenty of bug reports about it randomly dying. We got some boxes still running .32 simply because they haven't crashed.... yet.

    Francisco

  • yomeroyomero Member
    edited April 2012

    @quirkyquark said: Maybe he can share his secrets for getting the very latest .32 OVZ kernels running stably?

    Exactly :S

    @dmmcintyre3 said: still 6mb with mksh

    Didn't knew that shell :P If it can do autocompletion for files and commands in the path I am happy.

    @Francisco said: his users aren't as brutal as they could be :)

    LOL!
    Well, tbh the mine is almost doing nothing... :D

  • prometeusprometeus Member, Host Rep

    will be the climate and Italian food ;)

    Thanked by 1yomero
  • raindog308raindog308 Administrator, Veteran

    I wonder if at(1) takes less memory than cron. It would not be a drop-in replacement - you'd have to write the job so it reschedules itself every time

  • rethinkvpsrethinkvps Member

    If anyone would like to purchase a LEB just throw me an email at [email protected] and I may be able to hook you up with one.

  • quirkyquarkquirkyquark Member

    @rethinkvps: specs?

  • rethinkvpsrethinkvps Member

    @quirkyquark said: @rethinkvps: specs?

    >

    Throw me an email and I will see what I can do for you.

  • DoznDozn Member
    edited May 2012

    @KuJoe said: Yeah, one day in the BuyVM IRC channel me and another person tried to see how low we could go with a Debian 6 install, I forget who it was right now but they got it to 4MB or something with dropbear and something different than rsyslogd I think.

    EDIT: Just lowered it to 5MB with syslog-ng

    That was me ^^, 3MB while not logged into SSH haha

  • jcalebjcaleb Member

    @Dozn said: That was me ^^, 3MB while not logged into SSH haha

    care to share how its done?

  • DoznDozn Member
    edited May 2012

    @jcaleb said: @Dozn said: That was me ^^, 3MB while not logged into SSH haha

    care to share how its done?

    It was a while ago now, but iirc, I started with BuyVM Debian Minimal, then doing:
    1. apt-get purge rsyslog && apt-get install inetutils-syslogd
    2. Replace SSH with Dropbear - apt-get purge ssh openssh-server openssh-client openssh-blacklist openssh-blacklist-extra && apt-get install dropbear
    3. BASH with DASH - apt-get install dash && chsh
    /bin/dash
    rm -f /bin/sh
    ln -s /bin/dash /bin/sh

    Edit: You lose a lot of useful functionality by switching to DASH, so it's not recommended unless having the extra RAM is critical. You also lose sFTP by switching to dropbear, although there's a couple scripts that allow you to switch to dropbear without losing it, but that's outside of this scope.

    Thanked by 1jcaleb
  • jcalebjcaleb Member

    i have a similar script to setup a debian 6, but always getting 6mb after ssh-ing to the vps. will check again.

  • lioncolalioncola Member
    edited May 2012
    • replace init with runit linked with dietlibc? (I'm working on this with KVM, not sure if can do in openvz)
    • do everything through the serial console
    Thanked by 1jcaleb
Sign In or Register to comment.