Every 2 minutes exactally ...... help

[rpmws]

Hi Guys ..I need help :-( I have no hair left at all.

Did major upgrade last night:
Old was

RH7.1 - 2.4.2-2 (old Apache / old PHP, no suexec)

went to 2.4.18 new apache, php, went to sueexec.

I have several sueexec issues ..so far very easy to take care of them. Apache seems ok ..php is ok.

Now here is the weirdest problem.

Ever since the upgrade.

I loose nettraffic every 160-162 seconds exactally. It goes down for 3-45-60 seconds at a time. This happens when pinging from 10 locations, internally at noc and from here at office. Pinging other boxes at same exact time on same switch have no problem.
Whatever it is , it's every 2 minutes.

Now.. when kernel was compiled the first thing was IP tables problems not matching... now IP tables aren't even in Kernel. I don't use bandmin. I have commented (I think cron for bandmin) if I try to do anything with iptables I get masive problems with it not being there. Nothing I have done in 10 hours has made any difference what so ever in the 2 minute &dead& spot every 2 minutes. Loads are normal.. kill portsentry, start ps ..nothing helps ..rpm a new bind ...same thing. ran every possible fixscript I can find. Server seems fine ..just sleep every 2 minutes. for 3-30 seconds at a time. Actually time frame now is maybe 3-6 seconds more than earlier where it was 5-30 seconds down. Traffic is lower now. I have tailed every log i can find and nothing is telling me anything. cron log is only mailman (which is busted) but who cares right now (sueexec related).. dcpumon is in there but that is 5 minutes.

Any thought????????
Thanks in advance.

 

[feanor]

Perhaps your ethernet driver got hosed up slightly when your kernel was upgraded. Did you upgrade your kernel from source or rpm?

What do you have in dmesg .... anything weird? If network is cutting out semi-frequently you have got to have something in dmesg, unless your machine is getting flooded by someone evil.

netstat -an , and look for anything out of the ordinary.

Perhaps someone took the IP to your machine that is on the same network? A long shot, I know.

If this is indeed related to a RH kernel upgrade or iptables/chains being screwed, it is definitely your networking card or configuration.

that's my .02 until more details are posted here

 

[rpmws]

I didn't do the kernel so I don't know but I am 80% sure it was rpm.

I get this in top:
4511 root 9 0 1768 1704 1296 S 0.0 0.1 0:00 stunnel-3.22loc
4513 root 9 0 1440 1328 1176 S 0.0 0.1 0:00 stunnel-3.22loc
4515 cpanel 9 0 1440 1328 1176 S 0.0 0.1 0:00 stunnel-3.22loc
4517 cpanel 9 0 1648 1584 1296 S 0.0 0.1 0:00 stunnel-3.22loc
4519 cpanel 9 0 1440 1328 1176 S 0.0 0.1 0:00 stunnel-3.22loc

That just started lately ..(more than one)

On the netstat, well it's hard to say ..not an expert. I can't look at it right when it happens becuase it's down. It does seem to be all my IPs go down. Main and aliases same exact time.

 

[rpmws]

I just shut off 2 IPs that looked a little weird in netstat and it made no difference.

 

[rpmws]

I have also tried turning off every service in WHM and just ping the IPs from here. Same thing. It's not a service.

 

[rpmws]

Only thing in dmesg was written early today.. 10K total

root@box [/var/log]# tail -f /var/log/dmesg

IP Protocols: ICMP, UDP, TCP, IGMP
IP: routing cache hash table of 8192 buckets, 64Kbytes
TCP: Hash tables configured (established 262144 bind 65536)
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
ds: no socket drivers loaded!
VFS: Mounted root (ext2 filesystem) readonly.
Freeing unused kernel memory: 244k freed
Adding Swap: 512024k swap-space (priority -1)
Adding Swap: 512024k swap-space (priority -2)
md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27

 

[feanor]

what's in /var/log/messages, when this occurs?

or anywhere close to that timeframe.....

anything?

 

[rpmws]

you see this above ?
&ds: no socket drivers loaded!&

 

[rpmws]

nope ..almost nothing pops up in messages. One of the 10 things I have been tailing. It's just random stuff like normal.