Hi, i been trying to restore a backup of a server but it only stops working at "Performing other Apache configuration" it doesnt show any error, it only stops working at that stage. I have modified the Quotas assigned and im shure theres no related to quotas. But i have to say its on a new vps plan with 1024MB in RAM. I have restored many times the same server before in another VPS Provider with 768MB RAM without any problem.
Thanks in advance!
-Theres what i get in proc/user_beancounters, i hope this could help to know whats happening. Thanks!
Version: 2.5
uid resource held maxheld barrier limit failcnt
3256: kmemsize 18638134 119177216 536870912 536870912 0
lockedpages 0 8 131072 131072 0
privvmpages 184660 936746 9223372036854775807 9223372036854775807 0
shmpages 1189 262913 9223372036854775807 9223372036854775807 0
dummy 0 0 0 0 0
numproc 61 126 9223372036854775807 9223372036854775807 0
physpages 10775 262147 0 262144 10
vmguarpages 0 0 9223372036854775807 9223372036854775807 0
oomguarpages 51655 753465 9223372036854775807 9223372036854775807 4
numtcpsock 29 41 9223372036854775807 9223372036854775807 0
numflock 304 312 9223372036854775807 9223372036854775807 0
numpty 1 1 9223372036854775807 9223372036854775807 0
numsiginfo 0 66 9223372036854775807 9223372036854775807 0
tcpsndbuf 525512 791736 9223372036854775807 9223372036854775807 0
tcprcvbuf 475136 7505968 9223372036854775807 9223372036854775807 0
othersockbuf 187272 446704 9223372036854775807 9223372036854775807 0
dgramrcvbuf 0 30056 9223372036854775807 9223372036854775807 0
numothersock 181 211 9223372036854775807 9223372036854775807 0
Howdy,
The physpages and oomguarpages limits are showing failures (those are visible in the "failcnt" field).
If you attempt restoring your backup again, do those failcnt fields increase?
-Eric
Hi, thats right, i can see that it increased. I have tried to restore another backup and i get, do you know if theres something that I can do? Thanks again!
Version: 2.5 uid resource held maxheld barrier limit failcnt 3256: kmemsize 22608677 119177216 536870912 536870912 0 lockedpages 0 8 131072 131072 0 privvmpages 213176 944498 9223372036854775807 9223372036854775807 0 shmpages 1829 262913 9223372036854775807 9223372036854775807 0 dummy 0 0 0 0 0 numproc 69 129 9223372036854775807 9223372036854775807 0 physpages 29862 262147 0 262144 11 vmguarpages 0 0 9223372036854775807 9223372036854775807 0 oomguarpages 62148 753465 9223372036854775807 9223372036854775807 5 numtcpsock 34 41 9223372036854775807 9223372036854775807 0 numflock 304 312 9223372036854775807 9223372036854775807 0 numpty 2 2 9223372036854775807 9223372036854775807 0 numsiginfo 0 69 9223372036854775807 9223372036854775807 0 tcpsndbuf 600952 992536 9223372036854775807 9223372036854775807 0 tcprcvbuf 557056 7505968 9223372036854775807 9223372036854775807 0 othersockbuf 190608 458264 9223372036854775807 9223372036854775807 0 dgramrcvbuf 0 30056 9223372036854775807 9223372036854775807 0 numothersock 185 230 9223372036854775807 9223372036854775807 0 dcachesize 6385121 101839192 268435456 268435456 0 numfile 1713 1990 9223372036854775807 9223372036854775807 0 dummy 0 0 0 0 0 dummy 0 0 0 0 0 dummy 0 0 0 0 0 numiptent 41 41 9223372036854775807 9223372036854775807 0
I have the same problem. And actually there is an error which is shown in the terminal of the corresponding machine - see my attachment.
Importing the backup a second time doesn't work as there is a user conflict since it seems the first try already added a lot of stuff. In my case that is not so curcial as I tried to import it in a fresh VM and can just jump back to the last snapshot but for other users this could get annoying.
By the way, the domain which should be imported is ~800 MB packed and the server which gave the oom error has 4 GB ram assigned.
Any fix or temporary workaround would be very much appreciated as it is a little urgent to move this domain in the next days (better sooner than later) for me (:
Best regards from Germany, Wutar
There is a bug where restoring a domain that has SSL enabled can eat up more memory than it should.
That bug is documented here, and there is a fix planned for that tomorrow (Saturday):
http://www.virtualmin.com/node/24523
What Wutar is describing may be that bug; I'm not so sure about dakser, you may just be seeing an OpenVZ resource limitation :-)
But, if the domain you're restoring has SSL enabled -- try restoring all your domain except for that one domain. And then, when the bugfix comes out on Saturday, apply that, and see if you're then able to restore the domain that hasn't been working.
-Eric
Great, thanks for the fast feedback!
In my case SSL is of course enabled so this can very well be the case and I'll try with the new release asap.
If I don't post again consider this issue as resolved and the fix to work ;-)
Regards, Wutar
I been trying to restore only SSL sites and then was when it stoped working in "Performing other Apache configuration" and showing in Virtualmin Panel that the use of memory is full. That happens with sites from even 100 MB, but i tested restoring a site without SSL of 700MB and it have been restored without any problem.
Thanks so much.
I just wanted to confirm that this issue happens to me as well on OpenVZ with Ubuntu 10.04 (virtualmin-base is version 1.0-29). I tried this on 2012-12-16 at 1:45PM EST. I'm not sure if the updated package with the fix has been pushed already, or if this is an OpenVZ issue. The restore that fails is also an SSL site (and is only ~60MB). We look forward to updates!
Thanks!
The Virtualmin 3.97-2 update is out now -- so try that restore again when you have a chance, and see if that works for you.
-Eric
Hi! Thanks a lot, with the update it works without any problem.