I tried to SSH into one of my VPSes and it failed (connection reset). So I logged in to the HN and tried enter, that failed. Usually when that happens it's "cannot fork" due to inadequate resource, but this time...
[root@sakura ~]# vzctl enter 112
enter failed
I checked UBC just in case...
[root@sakura ~]# vzctl exec 112 cat /proc/user_beancounters
Version: 2.5
uid resource held maxheld barrier limit failcnt
112: kmemsize 2936930 4623186 16384000 18022400 0
lockedpages 0 0 32 32 0
privvmpages 47892 245925 262144 288358 0
shmpages 140 812 8192 8192 0
dummy 0 0 0 0 0
numproc 52 80 100 100 0
physpages 12146 27743 0 2147483647 0
vmguarpages 0 0 6144 2147483647 0
oomguarpages 12146 27743 6144 2147483647 0
numtcpsock 10 18 80 80 0
numflock 6 12 100 110 0
numpty 0 1 16 16 0
numsiginfo 0 10 256 256 0
tcpsndbuf 101728 291600 5365760 10485760 0
tcprcvbuf 214120 507292 5365760 10485760 0
othersockbuf 22360 45640 1503232 4063232 0
dgramrcvbuf 0 8380 2097152 2097152 0
numothersock 18 28 80 80 0
dcachesize 0 0 1048576 1097728 0
numfile 1164 1706 2048 2048 0
dummy 0 0 0 0 0
dummy 0 0 0 0 0
dummy 0 0 0 0 0
numiptent 10 10 128 128 0
I tried restarting it, and it still had the same problem. I checked dmesg and noticed:
UBC: BUG: fixing bad page PBs for c1646fbc
flags: 0x20000070 mapping: ef9c43fd mapcount: 0 count: 1
PB is c2e56da0 magic 1651769345
[<c015f2ab>] pb_remove_ref+0x34b/0x460
[<c016b380>] find_get_page+0x50/0x130
[<c018133a>] zap_pte_range+0x18a/0x2e0
[<c015ec88>] pb_add_ref+0x1f8/0x4d0
[<c0181513>] zap_pmd_range+0x83/0xb0
[<c01816bf>] unmap_page_range+0x5f/0xf0
[<c0181870>] unmap_vmas+0x120/0x200
[<c0189105>] exit_mmap+0xa5/0x260
[<c012d0ce>] mmput+0x7e/0x130
[<c01356ee>] do_exit+0x1be/0x860
[<c0135ece>] do_group_exit+0xbe/0x1d0
[<c0120c40>] do_page_fault+0x0/0x5dd
[<c057b61f>] syscall_call+0x7/0xb
Leak PB c2e56da0 for UB f6af0580
removed 1 PBs from hash
VPS: 112: stopped
VPS: 112: started
Registered tone zone 0 (United States / North America)
The stopped/started was me trying to restart it. It runs asterisk with ztdummy.
1. What does that UBC BUG mean?
2. Would ztdummy have something to do with it?
I have never had this issue before.
Off-topic:
Is there a specific reason why ipt_owner isn't built into the OVZ stab kernel? Is it because it needs to be virtualized to work properly?
[root@sakura ~]# uname -ar
Linux sakura.seraphimnetworks.com 2.6.9-023stab032.1 #1 Fri Oct 20 02:48:36 MSD 2006 i686 i686 i386 GNU/Linux
About the only thing I haven't tried yet is restarting the HN (I kinda want to avoid doing that if possible...)