Continued Issues with VPS

[nickbits]

I went to update a thread of mine today to find it close. The thread http://www.eukhost.com/forums/f29/vp...html#post85019 details some of the issues I have been having, although the thread was taken over by someone else.

Anyway, getting to the point, you guys (EUKHost), after pointless discussions on livechat, the guys on the forums did put things into perspective and actually gave me some proper answers as well as offering a few suggestions as to what the problem was. This I do appreciate. I am still looking at this, however I have also notice the server/sites are going down on a daily basis for the last several days, before that it was fine for a few days. However the is not much difference in the visitors to the sites, not that I can see anyway.

Getting to the point, after reading another thread http://www.eukhost.com/forums/f29/vp...owntime-14078/ who appeared to be experienceing similar issues, you mentioned that Virtuozzo is an issue. Before I continue to investigate other problems, could this be what is causing my issues?

It is just I cannot see what is causing the issues at present, although as I said I am still investigating the cause. However if I continue to draw a blank, I will have to look at the possibilty of moving servers/services to another company.

Regards,
Nick

[Rock]

I went to update a thread of mine today to find it close. The thread http://www.eukhost.com/forums/f29/vp...html#post85019 details some of the issues I have been having, although the thread was taken over by someone else.

Anyway, getting to the point, you guys (EUKHost), after pointless discussions on livechat, the guys on the forums did put things into perspective and actually gave me some proper answers as well as offering a few suggestions as to what the problem was. This I do appreciate. I am still looking at this, however I have also notice the server/sites are going down on a daily basis for the last several days, before that it was fine for a few days. However the is not much difference in the visitors to the sites, not that I can see anyway.

Getting to the point, after reading another thread http://www.eukhost.com/forums/f29/vp...owntime-14078/ who appeared to be experienceing similar issues, you mentioned that Virtuozzo is an issue. Before I continue to investigate other problems, could this be what is causing my issues?

It is just I cannot see what is causing the issues at present, although as I said I am still investigating the cause. However if I continue to draw a blank, I will have to look at the possibilty of moving servers/services to another company.

Regards,
Nick

Hi Nick,

If it's really an issue with Virtuozzo (which might be true in this case), I'd recommend having your Linux-VPS-01 moved to our VMWare cloud.

A Linux VPS 01 package comes with 512MB RAM, which is the cause of the problem you're facing. Please find the attached kernel logs from the host server clearly displaying the processes being killed due to lack of memory. OOM is Out of Memory & 16810 is the ID of your VPS.

kill_signal(16810.2): selecting to kill, queued 0, seq 7066, exc 828 0 goal 3707 0...
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7066, exc 3707 0 red 54033 865
kill_signal(16810.2): selecting to kill, queued 0, seq 7068, exc 2733 0 goal 21773 0...
kill_signal(16810.2): task ffff8101f5a162a0, thg ffff81014f415500, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7068, exc 21773 0 red 54033 865
kill_signal(16810.2): task ffff8101f5a162a0, thg ffff81014f415500, sig 2
kill_signal(16810.2): selecting to kill, queued 0, seq 7069, exc 6705 0 goal 6705 0...
kill_signal(16810.2): task ffff81013cbca420, thg ffff8101762c6bc0, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7069, exc 6705 0 red 55226 868
kill_signal(16810.2): task ffff81013cbca420, thg ffff8101762c6bc0, sig 2
kill_signal(16810.2): selecting to kill, queued 0, seq 7070, exc 2199 0 goal 165474 0...
kill_signal(16810.2): selected 4, signalled 4, queued 4, seq 7070, exc 165474 0 red 216564 3463
nofork(16810.2): delayed task ffff8101a1e9b1a0 for 301 msec
kill_signal(16810.2): selecting to kill, queued 0, seq 7071, exc 2812 0 goal 2812 0...
kill_signal(16810.2): task ffff810009674b20, thg ffff81003ef681c0, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7071, exc 2812 0 red 54032 865
kill_signal(16810.2): task ffff810009674b20, thg ffff81003ef681c0, sig 2
nofork(16810.2): delayed task ffff81006721ade0 for 562 msec
kill_signal(16810.2): selecting to kill, queued 0, seq 7072, exc 176 0 goal 13244 0...
kill_signal(16810.2): task ffff8101b26725e0, thg ffff81002b795840, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7072, exc 13244 0 red 54031 865
kill_signal(16810.2): task ffff8101b26725e0, thg ffff81002b795840, sig 2
kill_signal(16810.2): selecting to kill, queued 0, seq 7074, exc 4153 0 goal 253212 0...
kill_signal(16810.2): selected 5, signalled 5, queued 5, seq 7074, exc 253212 0 red 268595 4325
kill_signal(16810.2): selecting to kill, queued 0, seq 7075, exc 3274 0 goal 3274 0...
kill_signal(16810.2): task ffff81012b446360, thg ffff8101630178c0, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7075, exc 3274 0 red 58070 868
kill_signal(16810.2): task ffff81012b446360, thg ffff8101630178c0, sig 2
kill_signal(16810.2): selecting to kill, queued 0, seq 7076, exc 1613 0 goal 1613 0...
kill_signal(16810.2): task ffff810140550aa0, thg ffff810087003880, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7076, exc 1613 0 red 54033 865
kill_signal(16810.2): task ffff810140550aa0, thg ffff810087003880, sig 2
nofork(16810.2): delayed task ffff81006ba015a0 for 562 msec
kill_signal(16810.2): selecting to kill, queued 0, seq 7077, exc 591 0 goal 591 0...
kill_signal(16810.2): task ffff81008b6d8d20, thg ffff8101a12c7b00, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7077, exc 591 0 red 58457 868
kill_signal(16810.2): task ffff81008b6d8d20, thg ffff8101a12c7b00, sig 2
kill_signal(16810.2): selecting to kill, queued 0, seq 7078, exc 2487 0 goal 2487 0...
kill_signal(16810.2): task ffff8101ea6f76a0, thg ffff8101cb66e880, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7078, exc 2487 0 red 57383 868
kill_signal(16810.2): task ffff8101ea6f76a0, thg ffff8101cb66e880, sig 2
kill_signal(16810.2): task ffff8101ea6f76a0, thg ffff8101cb66e880, sig 3
OOM killed process php (pid=29580, ve=16810) exited, free=33888 gen=28604.
kill_signal(16810.2): selecting to kill, queued 0, seq 7079, exc 1568 0 goal 31558 0...
kill_signal(16810.2): task ffff810053756a20, thg ffff8101c3030500, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7079, exc 31558 0 red 57169 868
kill_signal(16810.2): selecting to kill, queued 0, seq 7080, exc 176 0 goal 1723 0...
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7080, exc 1723 0 red 54031 865
OOM killed process exim (pid=16810, ve=16820) exited, free=37240 gen=28659.
kill_signal(16810.2): selecting to kill, queued 0, seq 7081, exc 2263 0 goal 2263 0...
kill_signal(16810.2): task ffff81008a735260, thg ffff810043d7fc40, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7081, exc 2263 0 red 57959 868
nofork(16810.2): delayed task ffff810213a629a0 for 621 msec
nofork(16810.2): delayed task ffff8101de908360 for 983 msec
kill_signal(16810.2): selecting to kill, queued 0, seq 7082, exc 4523 0 goal 256459 0...
kill_signal(16810.2): selected 6, signalled 6, queued 6, seq 7082, exc 256459 0 red 270431 4932
kill_signal(16810.2): selecting to kill, queued 0, seq 7083, exc 6153 0 goal 216929 0...
kill_signal(16810.2): selected 4, signalled 4, queued 4, seq 7083, exc 216929 0 red 222071 3460
kill_signal(16810.2): selecting to kill, queued 0, seq 7084, exc 3018 0 goal 20538 0...
kill_signal(16810.2): task ffff8101a0e9aa20, thg ffff810043d7fc40, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7084, exc 20538 0 red 55634 865
nofork(16810.2): delayed task ffff81001b0400a0 for 4096 msec
kill_signal(16810.2): selecting to kill, queued 0, seq 7085, exc 3286 0 goal 8531 0...
kill_signal(16810.2): task ffff81006c396260, thg ffff8102144357c0, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7085, exc 8531 0 red 54488 865
kill_signal(16810.2): task ffff81006c396260, thg ffff8102144357c0, sig 2
kill_signal(16810.2): selecting to kill, queued 0, seq 7087, exc 1390 0 goal 74149 0...
kill_signal(16810.2): selected 2, signalled 2, queued 2, seq 7087, exc 74149 0 red 103436 1730
kill_signal(16810.2): selecting to kill, queued 0, seq 7088, exc 170 0 goal 15017 0...
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7088, exc 15017 0 red 49649 865
kill_signal(16810.2): selecting to kill, queued 0, seq 7094, exc 9582 0 goal 188657 0...
kill_signal(16810.2): selected 4, signalled 4, queued 4, seq 7094, exc 188657 0 red 189984 3445
kill_signal(16810.2): selecting to kill, queued 0, seq 7097, exc 1022 0 goal 1022 0...
kill_signal(16810.2): task ffff8101193436a0, thg ffff81022bc82100, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7097, exc 1022 0 red 42308 805
kill_signal(16810.2): selecting to kill, queued 0, seq 7098, exc 871 0 goal 871 0...
kill_signal(16810.2): task ffff8100248ff4e0, thg ffff8101eb110800, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7098, exc 871 0 red 55076 865
kill_signal(16810.2): task ffff8100248ff4e0, thg ffff8101eb110800, sig 2
kill_signal(16810.2): selecting to kill, queued 0, seq 7099, exc 8427 0 goal 634131 0...
kill_signal(16810.2): selected 18, signalled 18, queued 18, seq 7099, exc 634131 0 red 591506 12741
OOM killed process httpd (pid=703, ve=16810) exited, free=15301 gen=28868.
OOM killed process php (pid=4803, ve=16810) exited, free=39140 gen=28869.
OOM killed process php (pid=4806, ve=16810) exited, free=44844 gen=28870.
OOM killed process php (pid=4804, ve=16810) exited, free=46733 gen=28871.
nofork(16810.2): delayed task ffff810213a629a0 for 372 msec
OOM killed process leechprotect (pid=18443, ve=16810) exited, free=48781 gen=28872.
OOM killed process httpd (pid=25024, ve=16810) exited, free=46673 gen=28873.
nofork(16810.2): delayed task ffff810040558c60 for 673 msec
OOM killed process httpd (pid=22688, ve=16810) exited, free=59571 gen=28874.
OOM killed process php (pid=4825, ve=16810) exited, free=66515 gen=28875.
OOM killed process httpd (pid=711, ve=16810) exited, free=73521 gen=28876.
OOM killed process php (pid=4823, ve=16810) exited, free=77987 gen=28877.
OOM killed process php (pid=4802, ve=16810) exited, free=77987 gen=28878.
OOM killed process httpd (pid=18454, ve=16810) exited, free=78297 gen=28879.
OOM killed process httpd (pid=22695, ve=16810) exited, free=78483 gen=28880.
OOM killed process httpd (pid=4824, ve=16810) exited, free=78607 gen=28881.
OOM killed process httpd (pid=22987, ve=16810) exited, free=78824 gen=28882.
OOM killed process httpd (pid=4805, ve=16810) exited, free=78793 gen=28883.
OOM killed process php (pid=4800, ve=16810) exited, free=88713 gen=28884.
OOM killed process httpd (pid=4822, ve=16810) exited, free=88713 gen=28885.
kill_signal(16810.2): selecting to kill, queued 0, seq 7100, exc 6577 0 goal 6577 0...
kill_signal(16810.2): task ffff81016f433620, thg ffff810174d0c440, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7100, exc 6577 0 red 55808 865
kill_signal(16810.2): task ffff81016f433620, thg ffff810174d0c440, sig 2
kill_signal(16810.2): selecting to kill, queued 0, seq 7101, exc 1935 0 goal 145608 0...
kill_signal(16810.2): selected 3, signalled 3, queued 3, seq 7101, exc 145608 0 red 166924 2598
nofork(16810.2): delayed task ffff8101cb4540a0 for 403 msec
nofork(16810.2): delayed task ffff8101bad98a60 for 1340 msec
kill_signal(16810.2): selecting to kill, queued 0, seq 7102, exc 2776 0 goal 86781 0...
kill_signal(16810.2): task ffff810021461320, thg ffff810191229100, sig 1
kill_signal(16810.2): task ffff810053e8ac20, thg ffff8101e0cf0100, sig 1
kill_signal(16810.2): selected 2, signalled 2, queued 2, seq 7102, exc 86781 0 red 116069 1736
kill_signal(16810.2): selecting to kill, queued 0, seq 7103, exc 643 0 goal 51916 0...
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7103, exc 51916 0 red 54630 865
kill_signal(16810.2): task ffff8102272182e0, thg ffff810186d84540, sig 2
nofork(16810.2): delayed task ffff81013cc09460 for 3691 msec
kill_signal(16810.2): selecting to kill, queued 0, seq 7105, exc 6437 0 goal 202974 0...
kill_signal(16810.2): selected 4, signalled 4, queued 4, seq 7105, exc 202974 0 red 208477 3460
kill_signal(16810.2): selecting to kill, queued 0, seq 7106, exc 428 0 goal 25353 0...
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7106, exc 25353 0 red 53653 865
kill_signal(16810.2): selecting to kill, queued 0, seq 7107, exc 972 0 goal 33350 0...
kill_signal(16810.2): task ffff8102158301a0, thg ffff8101df1f3180, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7107, exc 33350 0 red 53318 865
nofork(16810.2): delayed task ffff8100c17db620 for 4569 msec
kill_signal(16810.2): selecting to kill, queued 0, seq 7108, exc 1261 0 goal 105391 0...
kill_signal(16810.2): selected 3, signalled 3, queued 3, seq 7108, exc 105391 0 red 134165 2565
kill_signal(16810.2): selecting to kill, queued 0, seq 7109, exc 2762 0 goal 264860 0...
kill_signal(16810.2): selected 7, signalled 7, queued 7, seq 7109, exc 264860 0 red 285151 5830
kill_signal(16810.2): selecting to kill, queued 0, seq 7110, exc 6356 0 goal 394584 0...
kill_signal(16810.2): selected 10, signalled 10, queued 10, seq 7110, exc 394584 0 red 422031 8347
OOM killed process php (pid=2607, ve=16810) exited, free=34626 gen=28886.
OOM killed process php (pid=2683, ve=16810) exited, free=42004 gen=28887.
OOM killed process php (pid=2660, ve=16810) exited, free=46063 gen=28888.
OOM killed process php (pid=2707, ve=16810) exited, free=46776 gen=28889.
OOM killed process php (pid=2602, ve=16810) exited, free=52232 gen=28890.
OOM killed process php (pid=2622, ve=16810) exited, free=61348 gen=28891.
OOM killed process php (pid=2386, ve=16810) exited, free=61937 gen=28892.
OOM killed process php (pid=2700, ve=16810) exited, free=62042 gen=28893.
OOM killed process php (pid=2387, ve=16810) exited, free=72303 gen=28894.
OOM killed process php (pid=2709, ve=16810) exited, free=50462 gen=28895.
kill_signal(16810.2): selecting to kill, queued 0, seq 7111, exc 2627 0 goal 35457 0...
kill_signal(16810.2): task ffff8100cf794b60, thg ffff81016d45c8c0, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7111, exc 35457 0 red 54310 865
kill_signal(16810.2): task ffff8100cf794b60, thg ffff81016d45c8c0, sig 2
kill_signal(16810.2): selecting to kill, queued 0, seq 7114, exc 4076 0 goal 136349 0...
kill_signal(16810.2): selected 3, signalled 3, queued 3, seq 7114, exc 136349 0 red 148722 2595
kill_signal(16810.2): selecting to kill, queued 0, seq 7117, exc 1039 0 goal 102819 0...
kill_signal(16810.2): selected 3, signalled 3, queued 3, seq 7117, exc 102819 0 red 119017 2505
kill_signal(16810.2): selecting to kill, queued 0, seq 7119, exc 4504 0 goal 4504 0...
kill_signal(16810.2): task ffff81022031cda0, thg ffff810115a21540, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7119, exc 4504 0 red 59067 868
kill_signal(16810.2): task ffff81022031cda0, thg ffff810115a21540, sig 2
nofork(16810.2): delayed task ffff8101cb4540a0 for 382 msec
nofork(16810.2): delayed task ffff81003ad3eb60 for 1034 msec
kill_signal(16810.2): selecting to kill, queued 0, seq 7120, exc 1490 0 goal 1490 0...
kill_signal(16810.2): task ffff81019391ece0, thg ffff81009f109140, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7120, exc 1490 0 red 54151 865
kill_signal(16810.2): selecting to kill, queued 0, seq 7121, exc 1156 0 goal 10780 0...
kill_signal(16810.2): task ffff81019391ece0, thg ffff81009f109140, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7121, exc 10780 0 red 41751 805
kill_signal(16810.2): task ffff81019391ece0, thg ffff81009f109140, sig 2
kill_signal(16810.2): selecting to kill, queued 0, seq 7123, exc 7473 0 goal 7473 0...
kill_signal(16810.2): task ffff81005d4702e0, thg ffff8101bedb5780, sig 1
kill_signal(16810.2): selected 1, signalled 1, queued 1, seq 7123, exc 7473 0 red 42127 805
nofork(16810.2): delayed task ffff8101cb4540a0 for 811 msec

Looking at these logs, it can be clearly seen that it's a memory issue, hence it's recommended to cut off these excessive usages or upgrade the plan there itself if you want to stay on Virtuozzo. You'd definitely see a drastic improvement once you move this VPS to a VMWare cloud. Please let know how you'd like to proceed.

[John]

Hello Nick,

Virtuozzo's biggest weakness is that it can't properly share I/O time between each VPS on the node which CAN lead to problems. It also has a similar limitation on CPU sharing. Symptoms of this are a seemingly high load time on a VPS without any processes showing up in the logs as being responsible. I can clearly see that this isn't the cause of your problems. It really is just memory.

If you were on a call with me now and I was listening to your visitor numbers, I'd be recommending a VM with 1 GB RAM. After looking at your site visitor traffic last week I was somewhat amazed you hadn't experienced problems previously. As I alluded to last week, it would only take a few things combining on your VPS to push you over your current limit. Be that a slight increase in traffic numbers at a certain time of day (more people coming in a morning and then it being quiet in the afternoon), someone attempting to brute force one of your passwords and updates running could each push you over your threshold.

I'd recommend that you do opt for an upgrade. I'd personally prefer it if you opted for a 1.5 Ghz Cloud with 1 GB RAM, as I see Virtuozzo as obsolete now (it will be replaced from our website in the next few weeks), but to be honest a RAM increase on your current VPS would likely also solve your problem.

In each case I'll offer this upgrade to you free for 2 months while you evaluate it. I know you are currently sceptical as to our reasoning, so I hope this offer will help tempt you into giving it a try.

Our top VPS engineers have looked over your current VPS and have all come to the same conclusion.

[nickbits]

Hi John/Rock,

Thank you both for the reply. I just wanted to double check before I looked any further. I ahve heard several people say about Virtuozzo being an issue, but I had not read anywhere what the problems were, hence the enquiry.

Virtuozzo's biggest weakness is that it can't properly share I/O time between each VPS on the node which CAN lead to problems. It also has a similar limitation on CPU sharing. Symptoms of this are a seemingly high load time on a VPS without any processes showing up in the logs as being responsible. I can clearly see that this isn't the cause of your problems. It really is just memory.

As for upgrade, I don't see the option you mention on the website. The Cloud Servers I see are 800MHz, and they are at £33 a month. At present it is the cost that is the reason I have not upgraded, especially as few of the sites actually bring in any money. Can RAM be bought as an extra on the VPS I am currently on?

As I said, I am still reviewing the sites to see if I can reduce the load any. There are a few options to try, using Varnish and Pressflow for example. However, and to be completly honest, what I am looking at doing is breaking the sites up. I have business, personal and society/communty sites there at present, so what i am looking at doing is moving some, or all.

Again, Rock, John, thank you for the help here, much appreciated.
Regards,
Nick

[John]

Hi Nick,

If you go into the order process of the Cloud it will give you some configurable options.

I'll have one of the techs check the currently RAM allocation on the node you are on to see if we can bump the memory up there without a migration.

[nickbits]

Hi Nick,

If you go into the order process of the Cloud it will give you some configurable options.

I'll have one of the techs check the currently RAM allocation on the node you are on to see if we can bump the memory up there without a migration.

Hi Joh,

I will take a look at the cloud options. Again to be honest, I am considering various options including your cloud and other companies, along with brekaing the sites up. I am going to take my time making that decision, I need to way up all the pros & cons including cost.

I am spending some time today looking at the sites themselves to see what I can do to lower the memory footprint, fingers crossed!

Once again thank you.
Nick

[Rock]

Hi Joh,

I will take a look at the cloud options. Again to be honest, I am considering various options including your cloud and other companies, along with brekaing the sites up. I am going to take my time making that decision, I need to way up all the pros & cons including cost.

I am spending some time today looking at the sites themselves to see what I can do to lower the memory footprint, fingers crossed!

Once again thank you.
Nick

Good luck with that Nick, I hope you come out with a viable solution

[nickbits]

Once again thank you for the help.

I have tried various things, I ahve altered various ports, I have lowered the footprint with the use of Pressflow, that caused other issues. I did manage to get the server to go over a week without a single issue and then problems again, then another week without any issues and then problems again. Seems periodic and I just can't pin it down. All use the same software so if one of my site is causing the issue so should the others, especially as the others have more visitors than the one causing the problems.

Anyway, John, I have looked at the Cloud option, for what you are recommending it would cost about £50 a month. That would be fine if I was bringing the money in or had a high volume of traffic, many of the sites are non-profit and I simply do not have £600 per year to spare.

What I am probably going to do is to move the offending site to another server. If that solves the problem then I know it was that site and can hopfully trace the issue down further, if the problems continue with the current EUKHost server then I don't know, probably move everything.

Regards,
Nick

[John]

Hello Nick,

What I can do is give you 2-3 hours of the time of one of our senior technicians next week who is also a php developer. He should be able to point you in the right direction.

[nickbits]

Hello Nick,

What I can do is give you 2-3 hours of the time of one of our senior technicians next week who is also a php developer. He should be able to point you in the right direction.

John,

Thank you for the offer, afraid I am away on holiday next week. Besides which, I am an experienced PHP developer myself. The website that appears to cause the issues is using a CMS, specifically Drupal. I have gone there for help, as well and I am also using Drupal on various other sites, none of which seem to cause a problem. I just don't get what is going on with this specific site. What I need from the logs is the exact page and script causing the problem, that is where I am having slight issues at the moment.

Currently digging through the CMS logs to see if I can track it down.

The only other thing I could think of doing was to alter htaccess, or the httpd.conf and tailor the KeepAlive, MaxConnections, etc. further, as that is the only other thing I can think of that would cause the problems (too many simultaneous connections)...

Thanks again,
Nick

[nickbits]

Hi,

I have now started to migrate some, well one site, to another server. I will do the rest as time permits and when I am happy with the new server.

Anyway, over the past week, last in August, I have been getting yet more issues. I have had email going down at least three times which has required me to manually restart exim, POP3 and IMAP services. The latest event being today (4th Sept). Not only that, on the 3rd Sept I had the whole server go dead, which was resolved via Live Chat after it had been down for over an hour.

I would like to know

1. IF anyone else is having similar issues, esp. with email? Wondering if it could be a mass of spam. But then again, why are the services, (Exim, pop and imap) not resart automatically?

2. What is going on?

I have raised a ticket, so will keep an eye out for a reply.

Nick

[nickbits]

So after a little more digging, the logs show that at almost every point when I get errors, the named service crashes. I don't know why. I assume that named crashing is affecting some of the other services.

So wuestion is, how do I trace back why named crashes? Can't seem to see anything in the logs to indicate why it crashed.

[John]

Hi,

Named is almost always the first service to die when a server runs out of memory. I'll check logs when I'm in the office tomorrow to see if this is true for your server, or, if you are in a rush, you can pop onto livechat and ask a support agent if there have been any memory limits reached in the past few days.

[John]

Hi Nick,

I see 2 memory related errors for the 3rd and also one on the 30th. Going back further, it looks like it happens once a week or so. Sound about right to you?

That time is still available as offered above, whenever you have time to take advantage of it.

[nickbits]

It is far more than once a week. I have had problems Friday (2nd), Sat (3rd) and Sunday (4th) as well as earlier in the week. It ahd only been once a week roughtly, but last week has been almost every day.

I have not changed any settings, either on the websites or server. As pointed out in the previous post, named is almost always shown in the logs as failing when the issue occurs.

I still do not see that it is my code, esp. when I am told on live support that it is another VPS on the node causing problems!

[John]

Ni Nick,

I can just tell you what I see and I can see you are running out of memory and coming close fairly frequently.

Re live chat, I've heard of that happening, if another VPS is hammering the I/O, but I'd be amazed if support are letting that continue. Please PM me your chat transcripts and I'll have a look at them.

I'm also not blaming your code, I'm just saying that your amount of RAM leaves very little room for any even slight burst requirements.

[nickbits]

John,

The lady on live chat did say the other node/VPS had been shut down. If it is memory again I will take another look, however, I have not really noticed named going before, I typically notice the old messages:

kill_signal(16810.2): selecting to kill, queued 0, seq 7099, exc 8427 0 goal 634131 0...
kill_signal(16810.2): selected 18, signalled 18, queued 18, seq 7099, exc 634131 0 red 591506 12741
OOM killed process httpd (pid=703, ve=16810) exited, free=15301 gen=28868.

But I have not noticed them this last week for the errors I have experienced and I have not noticed anything saying out of memory, well at least not when I am getting the main issues. I will see about setting a script up to monitor memory usage and log it.

The problem this last week has been email has gone down several times and not automatically restarted, but everything else has worked fine, ftp and httpd and mysql for example. Hence why I did not think it was a memory issue. It is only the last week that issues have escalated with email. I have had email down, the server down completly at least once.

HTTPD did go offline on Sunday, along with other services, but most automatically restarted, it was just email (exim, pop, imap) that did not fire up properly, which is what is getting to me.

Hence to me it would seem a different issue than before.

[nickbits]

Forgot to say, I know you are not blaming the code, and I know RAM has become an issue, hence the reason I have got another server. I will be offloading some of the sites to that server over the next month or two. That should ease the problem.

And I do appreciate the help, so again thank you.

[nickbits]

I asked on a ticket for this to be looked into. Got a reply with logs attached at 12.45 and was told the kernel was being updated on my VPS and that my server would be down for 10-15 mins.

First of, why is this done mid-day when my clients will want to be accessing their sites?

Secondly, now 2.15PM and I am on live chat to sort this out as the server still is not up and running.

John I do appreciate the help, but so far the problem just gets worse, I don't understand why as I am not actually changing any settings, etc. and the sites are not getting any increase in visitor numbers.

[nickbits]

So 30 mins on Live Chat and still not up...If it is a memory issue, can I not purchase additional memory until I move the sites elsewhere?