Re: Good and Bad Durons
quote:
--------------------------------------------------------------------------------
Originally posted by Ric
I consider myself very lucky that apparently I got one of the good Duron boxes. It has never had the problems that I see a lot of people complaining about and has been up for 35 days now. That would be 70 some days but I did have one crash when trying to compile PHP.
Rick
--------------------------------------------------------------------------------
Well, I apparently spoke to soon. My box just crashed again while compiling BitchX. All services down, had to TT a reboot.
Same thing happened last time I tried to do a make install on PHP so basically, I am now afraid to compile anything.
So what does that mean? If I can't get a binary that I trust then I am just SOL? Christ, I compile stuff all the time on our local redhat box and it is a P1 100... am pretty upset about this.
Rick
Full Version: Durons can't handle compiling?
I've never had a problem with my Duron box, including compiling PHP.
Same here with two Durons -- though I do recognize that there must be quite a few lemons among the Durons with the amount of similar problem people seems to have.
I'd probably recommend anyone planning to use a server in a production environment to run it in a couple of days under hard load, just to make sure there are no lose connections.
You can for exampel download seti@home (http://setiathome.ssl.berkeley.edu/download.html) and let it run for a couple of days. That'll catch any dodgy memory or heat problems and in the meantime you help finding ET
I'd probably recommend anyone planning to use a server in a production environment to run it in a couple of days under hard load, just to make sure there are no lose connections.
You can for exampel download seti@home (http://setiathome.ssl.berkeley.edu/download.html) and let it run for a couple of days. That'll catch any dodgy memory or heat problems and in the meantime you help finding ET
ya i called gave my 30 day notice today on my server since i got my webreseller box in my hands finally today
speeds are ownj on it
speeds are ownj on it
The box in general has been really solid and the ONLY time I have had problems that were not generated by my own inexperience has been these two times when compiling. All other things have just been little stuff I have been able to diagnose and fix.
Don't get me wrong, I am happy with the server and I am happy with RS. They have been very responsive on the few TT's I have had to open but I try not to bother them unless there is no alternative like a reboot. I am not the kind of person who runs off to promises of greener pastures, I like it here.
However, this compiling issue does bother me. I know it is not just me or my machine because I have read enough posts regarding this same problem. It seems verry odd, I mean, our little P100 with 32 MB of ram handles compilations that have taken an hour or more without skipping a beat so what could possibly be wrong with a much higher cpu/mem box doing it?
Rick
Don't get me wrong, I am happy with the server and I am happy with RS. They have been very responsive on the few TT's I have had to open but I try not to bother them unless there is no alternative like a reboot. I am not the kind of person who runs off to promises of greener pastures, I like it here.
However, this compiling issue does bother me. I know it is not just me or my machine because I have read enough posts regarding this same problem. It seems verry odd, I mean, our little P100 with 32 MB of ram handles compilations that have taken an hour or more without skipping a beat so what could possibly be wrong with a much higher cpu/mem box doing it?
Rick
It could be any number of things, a slightly overheating CPU, RAM that has a minor issue, circuit board problems, it's hard to say.
Right, that I know but it does not add up if you think about it. None of those problems are going to be present just occasionally or at least, it would be unusual.
I have had about a hundred computers and experienced all of those problems at one time or another but it was not an off and on thing. If ram is flukey there would be other indications and the same with bad boards. I have had cpu heat problems too but hot was hot and you didn't have to be number crunching to know there was a problem.
Never had an AMD though, I have exclusively bought Intel so that part I don't know anything about. Do these chips have a history of heat problems or critical temp ranges?
My gut feeling would tell me hard drive or drive interface more than anything else but then I would expect some kind of error or errors then crash instead of the thing just locking up in the middle of a compile without any warning.
Rick
I have had about a hundred computers and experienced all of those problems at one time or another but it was not an off and on thing. If ram is flukey there would be other indications and the same with bad boards. I have had cpu heat problems too but hot was hot and you didn't have to be number crunching to know there was a problem.
Never had an AMD though, I have exclusively bought Intel so that part I don't know anything about. Do these chips have a history of heat problems or critical temp ranges?
My gut feeling would tell me hard drive or drive interface more than anything else but then I would expect some kind of error or errors then crash instead of the thing just locking up in the middle of a compile without any warning.
Rick
Some types of errors never really show up unless there's heavy memory and processor usage, it's frequently hard to say. AMD chips do tend to run fairly hot, but it's not an issue with sufficient cooling.
Well, I don't know what I can do about it short of dumping the thing an getting a new one assuming what I want is even available. Not being able to compile from source turns the box into an antique software wise in about six months, I can't run an efficient, secure server bottom line.
I could try it again and crash it again and just keep breaking it until they tear it down and fix it. What a terrible thing to do to clients on that box, that is just not an option.
Oh and madsere, Been running seti@home since the beta... were still looking
Rick
I could try it again and crash it again and just keep breaking it until they tear it down and fix it. What a terrible thing to do to clients on that box, that is just not an option.
Oh and madsere, Been running seti@home since the beta... were still looking
Rick
If I had a lemon I'd probably just dump it and get one of those one buck Compaq jobs. It's just too much of a headache with broken hardware as is apparent from posts here. I consider myself lucky mine doen't seem to have any problems.
Well, we may have seen the end of the $1 specials. It was sweet.
Gary
Gary
I want to update this issue for anyone following it and make a note to people here who say RS is not responsive to TT's.
Our experience has been the opposite, they have always initially responded within minutes. This problem was no different and as you can see by the updated TT info below, they are not just doing a reboot and forgetting about it.
5/29/02 3:11:09 AM
pulling server to test ram
5/29/02 3:31:51 AM
bad data bits, and memory could not operate at full 133 mhz.
please swap memory
I don't think they were obligated to run diags the first time this happened with a PHP compile. I would not have jumped into diagnostics immediately if a problem like that occurred here with one of our computers either. I would have rebooted and started looking deeper into the problem if it happened a second time just as they are doing.
Personally I am happy as can be with RS and think it is only fair to praise them when they do the right thing. It is easy to bitch and I wonder how many people had their machines fixed and never even bothered to post an update to a complaint. I never asked them in the TT to do diags, I asked for a reboot and told them this is the second time it has happened out of two tries on a compile. They knew I was concerned about it but I did not threaten to go somewhere else or insult them, that is just counter productive.
New people... you may want to think about this real hard before you use your TT's rashly or start screaming instead of asking questions. If you open TT's on little stuff that you should be taking care of yourself, your going to lose credibility for when a real problem develops.
Nobody is perfect and I am sure RS has good techs and bad as well as trouble periods for whatever reasons. Angry words are not going to help move your TT along any faster, be sane.
Rick
Our experience has been the opposite, they have always initially responded within minutes. This problem was no different and as you can see by the updated TT info below, they are not just doing a reboot and forgetting about it.
5/29/02 3:11:09 AM
pulling server to test ram
5/29/02 3:31:51 AM
bad data bits, and memory could not operate at full 133 mhz.
please swap memory
I don't think they were obligated to run diags the first time this happened with a PHP compile. I would not have jumped into diagnostics immediately if a problem like that occurred here with one of our computers either. I would have rebooted and started looking deeper into the problem if it happened a second time just as they are doing.
Personally I am happy as can be with RS and think it is only fair to praise them when they do the right thing. It is easy to bitch and I wonder how many people had their machines fixed and never even bothered to post an update to a complaint. I never asked them in the TT to do diags, I asked for a reboot and told them this is the second time it has happened out of two tries on a compile. They knew I was concerned about it but I did not threaten to go somewhere else or insult them, that is just counter productive.
New people... you may want to think about this real hard before you use your TT's rashly or start screaming instead of asking questions. If you open TT's on little stuff that you should be taking care of yourself, your going to lose credibility for when a real problem develops.
Nobody is perfect and I am sure RS has good techs and bad as well as trouble periods for whatever reasons. Angry words are not going to help move your TT along any faster, be sane.
Rick
Excellent post Rick! I was one of those unlucky people with a bad box (I had to post a TT the very first day as RS mistakenly installed Ensim on my box on set up instead of Plesk - and for the first 5 weeks it went steadily downhill from there) but I have only returned once to post how cool they have been on other issues. They always take care of the problems quickly or at least jump in and try to help quickly. Even though I have been unlucky, I too am very pleased with RS overall. Been up for 24 days and counting, feels pretty good right now...
Let me first say the Duron I have now is rock solid, and if a computer crashes only when compiling, then something is not right.
I got my Duron and did got a little burn in program going at the same time I kept compiling the kernel, and it crashed. Rackshack quickly tested and replaced memory, but then it crashed differently, so they did a server swap, and since then rock solid. So if you are having problems it is not Duron's in general, but you got a bad one. Open a TT.
I got my Duron and did got a little burn in program going at the same time I kept compiling the kernel, and it crashed. Rackshack quickly tested and replaced memory, but then it crashed differently, so they did a server swap, and since then rock solid. So if you are having problems it is not Duron's in general, but you got a bad one. Open a TT.
Yup, my first duron was a lemon as well...took 5 days to finally get a new server, but this new one is rock solid.
Glad to hear RS is taking care of your server, Ric.
I'm glad I got one of the good Durons. For a couple of months there after reading others' complaints I held my breath every time I typed `make'. Now I don't even think about it, everything works great as long as I don't accidentally type a `chown -R' command and wreck Plesk again..
I'm glad I got one of the good Durons. For a couple of months there after reading others' complaints I held my breath every time I typed `make'. Now I don't even think about it, everything works great as long as I don't accidentally type a `chown -R' command and wreck Plesk again.
.
Thanks guys, I am just happy, happy, happy that it is fixed and intend to run some tests on it today and tonight... fingers crossed.
It pays to be up front with clients. As soon as I knew there was a problem I notified them of an impending short outage for mem swap and although I did not think any of them were experiencing web problems (because we were not) we wanted to keep them in the know. I got this email today from one of our bigger clients.
Hopefully everything will check out ok today and I will finally be able to get PHP upgraded. Then I can get on to more important things like installing that TFC game server. Going to have to find some good rcons though because I am far to mean to be trusted with that kind of power
It pays to be up front with clients. As soon as I knew there was a problem I notified them of an impending short outage for mem swap and although I did not think any of them were experiencing web problems (because we were not) we wanted to keep them in the know. I got this email today from one of our bigger clients.
QUOTE
Hey Rick!
Thanks for the notification and results. I was actually about to email you
prior to your first email, because we did notice a problem a couple days
before that. When I received your email about the problems I
breathed a sigh of relief, because I am hopeful that the issue will now be
resolved with the new RAM. Ok, now you're obviously very curious as to what
the problem we were having with that server was...
Basically, the files that are on the drive change (become corrupted) after
a period of time. One of our biggest and most popular files was uploaded and then downloaded and tested to be
fine. A couple of days later, some customers were complaining that they
were getting a corrupt install program. We tracked it down that they were
all using that particular server. We downloaded the file and verified that
it was indeed corrupt (remember that after first being uploaded it was
definitely fine). Anyway, we re-uploaded it, and tested it (downloaded
again), and it was fine again. Then, a few days later it was again corrupt!
So the issue is that files on the hard disk of the server (presumably files
which are frequently accessed), *can* and eventually *do* become corrupt.
This is obviously a serious problem for any computer, let alone a server.
Once I was sure it was the server, and had had the problem repeat itself, I
was going to email you.
Anyway, to my knowledge, the problem has not re-occured since you have
addressed the RAM problem. We will continue to monitor the downloads daily,
and let you know if the problem occurs again.
Thanks for taking the time to email us about the issues as you knew about
them.
Cheers,
Adam
Thanks for the notification and results. I was actually about to email you
prior to your first email, because we did notice a problem a couple days
before that. When I received your email about the problems I
breathed a sigh of relief, because I am hopeful that the issue will now be
resolved with the new RAM. Ok, now you're obviously very curious as to what
the problem we were having with that server was...
Basically, the files that are on the drive change (become corrupted) after
a period of time. One of our biggest and most popular files was uploaded and then downloaded and tested to be
fine. A couple of days later, some customers were complaining that they
were getting a corrupt install program. We tracked it down that they were
all using that particular server. We downloaded the file and verified that
it was indeed corrupt (remember that after first being uploaded it was
definitely fine). Anyway, we re-uploaded it, and tested it (downloaded
again), and it was fine again. Then, a few days later it was again corrupt!
So the issue is that files on the hard disk of the server (presumably files
which are frequently accessed), *can* and eventually *do* become corrupt.
This is obviously a serious problem for any computer, let alone a server.
Once I was sure it was the server, and had had the problem repeat itself, I
was going to email you.
Anyway, to my knowledge, the problem has not re-occured since you have
addressed the RAM problem. We will continue to monitor the downloads daily,
and let you know if the problem occurs again.
Thanks for taking the time to email us about the issues as you knew about
them.
Cheers,
Adam
Hopefully everything will check out ok today and I will finally be able to get PHP upgraded. Then I can get on to more important things like installing that TFC game server. Going to have to find some good rcons though because I am far to mean to be trusted with that kind of power
First tests on a compile crashed the server again, TT led to a test of the new RAM, no errors. Further diags show drive bad, must replace, total data loss.
Ok, no problem. However, I want to prepare so this will be as little work as possible so here is my plan.
1. Delete all log files.
2. Tarball the /usr/local/psa/* complete.
3. Tarball the /etc/* complete.
Those first three are just in case the following don't work and I have to restore config files or client data manually. I would also like to grab a complete SQL dump before I begin but I don't know how to do that, any suggestions?
According to the Plesk instructions on their new backup/restore utility...
Using BU for PSA upgrade from v.2.0.1 to v.2.5.1.
=================================================
It is possible to upgrade PSA from v.2.0.1 to v.2.5.1 by performing full
backup of PSA v.2.0.1 and its following restoring on freshly installed
PSA v.2.5.1 software or by performing simultaneous backup on the host
with PSA v.2.0.1 installed and restoring on another host with freshly
installed PSA v.2.5.1.
To have the PSA v.2.5.1 software installed, you should either install
RPM-version of PSA v.2.5.1 or install Standard version of PSA v.2.5.0
and then apply patch to it to get the PSA 2.5.1 version.
=================================================
So I should be able to perform a backup of all psa data with the util and then re-install on a fresh copy of Plesk on the new hard drive.
this is assuming I can get through a backup without crashing the server and if I can convince RS to install Plesk 2.5.1 on my new drive instead of the initial 2.0 that was installed.
Any suggestions on procedures or anything else I should backup before I start would be welcome!
Rick
Ok, no problem. However, I want to prepare so this will be as little work as possible so here is my plan.
1. Delete all log files.
2. Tarball the /usr/local/psa/* complete.
3. Tarball the /etc/* complete.
Those first three are just in case the following don't work and I have to restore config files or client data manually. I would also like to grab a complete SQL dump before I begin but I don't know how to do that, any suggestions?
According to the Plesk instructions on their new backup/restore utility...
Using BU for PSA upgrade from v.2.0.1 to v.2.5.1.
=================================================
It is possible to upgrade PSA from v.2.0.1 to v.2.5.1 by performing full
backup of PSA v.2.0.1 and its following restoring on freshly installed
PSA v.2.5.1 software or by performing simultaneous backup on the host
with PSA v.2.0.1 installed and restoring on another host with freshly
installed PSA v.2.5.1.
To have the PSA v.2.5.1 software installed, you should either install
RPM-version of PSA v.2.5.1 or install Standard version of PSA v.2.5.0
and then apply patch to it to get the PSA 2.5.1 version.
=================================================
So I should be able to perform a backup of all psa data with the util and then re-install on a fresh copy of Plesk on the new hard drive.
this is assuming I can get through a backup without crashing the server and if I can convince RS to install Plesk 2.5.1 on my new drive instead of the initial 2.0 that was installed.
Any suggestions on procedures or anything else I should backup before I start would be welcome!
Rick
I don't believe the problem is the disk.
You should try to install lm_sensors (though realizing that if compiles crash your server you'll have a problem installing it ... mayeb you have another identical server you could use for the compile?) and monitor temperature while you compile.
AMD Durons are notorious for running very hot. I measured over 60C on one of my AMD Durons during a test compile. It wouldn't take much of a worn out fan to bring that temperature up in a critical range; in fact I think already 60C is above what the Duron specs call for but so far I haven't had any trouble (knock on wood).
You should try to install lm_sensors (though realizing that if compiles crash your server you'll have a problem installing it ... mayeb you have another identical server you could use for the compile?) and monitor temperature while you compile.
AMD Durons are notorious for running very hot. I measured over 60C on one of my AMD Durons during a test compile. It wouldn't take much of a worn out fan to bring that temperature up in a critical range; in fact I think already 60C is above what the Duron specs call for but so far I haven't had any trouble (knock on wood).
Mad, here is the techs results...
I don't question your judgement on this, I think it is odd that the thing crashes instead of errors on compiling and that it only does it on a compile to begin with. However, since they have located problems on the hard disk, especially since corrupt files are a known problem with a clients data (see above posts), I think the hard drive should be replaced even if it is not the root cause of the problem. Do you agree?
rick
QUOTE
5/31/02 4:05:16 AM
checked for ram and it is good.
we are investigating the server for other problems.
5/31/02 8:30:17 AM
i have not noticed any memory errors in the logs . Currently I am test the hard disck to see if the problem resides there.
5/31/02 9:29:35 AM
Your hard drive does contain bad Block on the /partition
We Should replace the harddrive. This will reqiure down time, and you will be given a fresh file system
I am sorry to say, that this means you would lose all your data
Please let us know when we can go ahead with the hard drive swap.
Thank you.
checked for ram and it is good.
we are investigating the server for other problems.
5/31/02 8:30:17 AM
i have not noticed any memory errors in the logs . Currently I am test the hard disck to see if the problem resides there.
5/31/02 9:29:35 AM
Your hard drive does contain bad Block on the /partition
We Should replace the harddrive. This will reqiure down time, and you will be given a fresh file system
I am sorry to say, that this means you would lose all your data
Please let us know when we can go ahead with the hard drive swap.
Thank you.
I don't question your judgement on this, I think it is odd that the thing crashes instead of errors on compiling and that it only does it on a compile to begin with. However, since they have located problems on the hard disk, especially since corrupt files are a known problem with a clients data (see above posts), I think the hard drive should be replaced even if it is not the root cause of the problem. Do you agree?
rick
In my experience bad blocks create weird problems but crashes during compile as the only symptom does not sound likely.
If it was my personal disk I would probably just reformat it and install ext3 on it. I believe that cures most bad block problems. But in your case I agree, I'd probably just accept a new disk to eliminate this as a possible problem. I would also install ext3 as the first thing on the disk if it's not there already.
If it was my personal disk I would probably just reformat it and install ext3 on it. I believe that cures most bad block problems. But in your case I agree, I'd probably just accept a new disk to eliminate this as a possible problem. I would also install ext3 as the first thing on the disk if it's not there already.
What is ext3 and where can I get it?
The PSA backup was a piece of cake, very nice util. Hopefully that will resore all my domain data and I will just have to manually restore/re-install system configs and 3rd party software.
I will make another backup just before the drive swap and if RS won't install PSA 2.5.1 for me instead of the initial 2.0, I will do it myself before I do the restore.
Since I have to start over I intend to keep a detailed log of every step, primarily from a security standpoint. On the bright side, this will take a few days instead of the first time around when it took six weeks. Maybe others can benefit from my experience if I post the procedure.
Rick
The PSA backup was a piece of cake, very nice util. Hopefully that will resore all my domain data and I will just have to manually restore/re-install system configs and 3rd party software.
I will make another backup just before the drive swap and if RS won't install PSA 2.5.1 for me instead of the initial 2.0, I will do it myself before I do the restore.
Since I have to start over I intend to keep a detailed log of every step, primarily from a security standpoint. On the bright side, this will take a few days instead of the first time around when it took six weeks. Maybe others can benefit from my experience if I post the procedure.
Rick
ext3 is a journaled filesystem evolved from ext2 - the Linux standard. Other journaled filessystems include XFS and JFS for Linux, NTFS for Windrowse.
Journaled filesystems keep a journal of all disk writes so in theory you will never have a disk problem following a crash. It basically is hyper accounting for filesystems :-) Search for ext3 and you will find many discussions both about what it is, why it is good, how to see if it is on your system already, how to install it if it is not.... as always: Search is your friend
Journaled filesystems keep a journal of all disk writes so in theory you will never have a disk problem following a crash. It basically is hyper accounting for filesystems :-) Search for ext3 and you will find many discussions both about what it is, why it is good, how to see if it is on your system already, how to install it if it is not.... as always: Search is your friend
The hard drive did solve the problem or at least, I was able to get through makes tonight I was never able to get through before... not a twitch (knock on wood).
The restore was an initial nightmare, SQL/root password problems before I could even touch the thing that lasted for a couple days. RS support finally got it fixed after a number of TT's on different issues but in general, great job as always by the RS techs.
Once I had access it took me less than 12 hours to do what took me six weeks the first time, thanks to this forum more than anything else... what a great bunch of people.
Rick
The restore was an initial nightmare, SQL/root password problems before I could even touch the thing that lasted for a couple days. RS support finally got it fixed after a number of TT's on different issues but in general, great job as always by the RS techs.
Once I had access it took me less than 12 hours to do what took me six weeks the first time, thanks to this forum more than anything else... what a great bunch of people.
Rick
man glad to hear you having better luck than I ever did lol.
best of luck with the new hardware wish i would have been treated a bit differently while i was there but no big deal now i guess.
Zac
best of luck with the new hardware wish i would have been treated a bit differently while i was there but no big deal now i guess.
Zac
QUOTE
Originally posted by Ric
The hard drive did solve the problem or at least, I was able to get through makes tonight I was never able to get through before... not a twitch (knock on wood).
The restore was an initial nightmare, SQL/root password problems before I could even touch the thing that lasted for a couple days. RS support finally got it fixed after a number of TT's on different issues but in general, great job as always by the RS techs.
Once I had access it took me less than 12 hours to do what took me six weeks the first time, thanks to this forum more than anything else... what a great bunch of people.
Rick
The hard drive did solve the problem or at least, I was able to get through makes tonight I was never able to get through before... not a twitch (knock on wood).
The restore was an initial nightmare, SQL/root password problems before I could even touch the thing that lasted for a couple days. RS support finally got it fixed after a number of TT's on different issues but in general, great job as always by the RS techs.
Once I had access it took me less than 12 hours to do what took me six weeks the first time, thanks to this forum more than anything else... what a great bunch of people.
Rick
Ric, i am also having trouble compiling and make'ing with my AMD..
RS support has be OUTSTANDING during this time of trouble.. They tried to compile this:
Resolution Description:
6/29/02 6:35:08 PM
Server has froze during the Configure phase of a perl test compilation, please check ram.
stats before ./Configure -de:
total used free shared buffers cached
Mem: 496732 204704 292028 1200 39384 91604
-/+ buffers/cache: 73716 423016
Swap: 1020116 0 1020116
7:32pm up 28 min, 1 user, load average: 0.03, 0.01, 0.00
stats before ./Configure -de:
total used free shared buffers cached
Mem: 496732 204704 292028 1200 39384 91604
-/+ buffers/cache: 73716 423016
Swap: 1020116 0 1020116
7:32pm up 28 min, 1 user, load average: 0.03, 0.01, 0.00
The have responded to my TT mainly withing minutes.. i did have a long wait for one reboot, but support was actually doing a ./configure command for me.. (it crashed on me during the make though)
Im glad it crashed on them... so at least something will be fixed
( its the same satisfying feeling you get when u take your car to the mechanic and what you say is wrong happens to them too )
My server is junky but RS rocks!
RS support has be OUTSTANDING during this time of trouble.. They tried to compile this:
Resolution Description:
6/29/02 6:35:08 PM
Server has froze during the Configure phase of a perl test compilation, please check ram.
stats before ./Configure -de:
total used free shared buffers cached
Mem: 496732 204704 292028 1200 39384 91604
-/+ buffers/cache: 73716 423016
Swap: 1020116 0 1020116
7:32pm up 28 min, 1 user, load average: 0.03, 0.01, 0.00
stats before ./Configure -de:
total used free shared buffers cached
Mem: 496732 204704 292028 1200 39384 91604
-/+ buffers/cache: 73716 423016
Swap: 1020116 0 1020116
7:32pm up 28 min, 1 user, load average: 0.03, 0.01, 0.00
The have responded to my TT mainly withing minutes.. i did have a long wait for one reboot, but support was actually doing a ./configure command for me.. (it crashed on me during the make though)
Im glad it crashed on them... so at least something will be fixed
( its the same satisfying feeling you get when u take your car to the mechanic and what you say is wrong happens to them too )
My server is junky but RS rocks!
QUOTE
Originally posted by Ric
I want to update this issue for anyone following it and make a note to people here who say RS is not responsive to TT's.
Our experience has been the opposite, they have always initially responded within minutes. This problem was no different and as you can see by the updated TT info below, they are not just doing a reboot and forgetting about it.
5/29/02 3:11:09 AM
pulling server to test ram
5/29/02 3:31:51 AM
bad data bits, and memory could not operate at full 133 mhz.
please swap memory
I want to update this issue for anyone following it and make a note to people here who say RS is not responsive to TT's.
Our experience has been the opposite, they have always initially responded within minutes. This problem was no different and as you can see by the updated TT info below, they are not just doing a reboot and forgetting about it.
5/29/02 3:11:09 AM
pulling server to test ram
5/29/02 3:31:51 AM
bad data bits, and memory could not operate at full 133 mhz.
please swap memory
They replaced our drive and memory, since then it has been stable and I have had only one crash on a compile out of maybe 20, some of them were pretty heavy so that part seems to be fixed.
But... lol, I still have a serious problem with the machine because it simply will not handle anything that takes even a minor sustained amount of CPU time.
I tried running a game server and it consistently crashes after a few hours even when I reduced slots down to a minimal level. I was monitoring the thing when it crashed and the CPU resources were at 80-90% free.
To be honest, I think the AMD we got was just a lemon. There are people here with the exact same box and config who have no problem at all running multiple game servers so I know some of them, probably most of them are fine.
It runs standard services like a timex but average load on it is nil. I am hoping we can afford to get an Intel/Plesk machine soon so we can have a backup for this one. Once in place, we will dump this one, we probably will still buy a second but I want to be able to depend on both of them.
RS in general has been great through this, if I whined or continued to crash the thing I am confident that they would continue to tear into it until the problem was solved but that is not an option. I have about 30 clients on this box and I won't subject them to that, not while it runs fine if I don't push it.
Good luck, I hope you get it fixed. I would be interested to know how it the problem was resolved once you get to that point.
Rick
But... lol, I still have a serious problem with the machine because it simply will not handle anything that takes even a minor sustained amount of CPU time.
I tried running a game server and it consistently crashes after a few hours even when I reduced slots down to a minimal level. I was monitoring the thing when it crashed and the CPU resources were at 80-90% free.
To be honest, I think the AMD we got was just a lemon. There are people here with the exact same box and config who have no problem at all running multiple game servers so I know some of them, probably most of them are fine.
It runs standard services like a timex but average load on it is nil. I am hoping we can afford to get an Intel/Plesk machine soon so we can have a backup for this one. Once in place, we will dump this one, we probably will still buy a second but I want to be able to depend on both of them.
RS in general has been great through this, if I whined or continued to crash the thing I am confident that they would continue to tear into it until the problem was solved but that is not an option. I have about 30 clients on this box and I won't subject them to that, not while it runs fine if I don't push it.
Good luck, I hope you get it fixed. I would be interested to know how it the problem was resolved once you get to that point.
Rick
Well, the ram check turned out ok. They sent the ticket back out for further investigation.. They do know somthing is not right since it crashed on them also.. I need the machine to be somewhat stable.. there are no sites on it now.. but there will be around 20 or 30 once i know the machine can stay up for a day without crashing... my celery box i was able to do all these update and more with no sweat.... maybe win98 is on this server by mistake? heheheh 
QUOTE
Originally posted by foggy
maybe win98 is on this server by mistake? heheheh
maybe win98 is on this server by mistake? heheheh
Thats funny Froggy, thanks for the chuckle.
I would expect to be able to run 150-200 average client domains on this thing as it is without a problem, low bandwidth sites anyway.
Web/mail/ftp/etc services don't load the thing up. Just can't run game servers and would probably have a problem with high traffic PHP based sites and the like.
If I am careful to only put low bandwidth sites on the thing it would *probably* be ok but that is not acceptable in my opinion. I would not attempt to load this thing up to the point where it was using 50% of resources regardless of if this problem existed or not. It is a PC server after all, I don't expect it to run like a Sun.
Rick
Well, i cant install image magick either.. Server froze again during the make.
My trouble ticket that i opened 25 hours ago is still open, i guess they are still investigating... Ram check was ok...
I dont want to get testy.. but this is getting frustrating.... I have a server that can only do one thing well.... crash..
My trouble ticket that i opened 25 hours ago is still open, i guess they are still investigating... Ram check was ok...
I dont want to get testy.. but this is getting frustrating.... I have a server that can only do one thing well.... crash..
For those of you who have systems freeze up on you during heavy load I would suggest installing MRTG (follow Mouses howto, it's on this forum) and having it monitor your cpu/mb temperature. It won't preent the system from shutting down following heat problems, but if it is indeed that, at least you'll know what the problem is (and have some leverage against asking RS for a new fan)
QUOTE
Originally posted by madsere
For those of you who have systems freeze up on you during heavy load I would suggest installing MRTG (follow Mouses howto, it's on this forum) and having it monitor your cpu/mb temperature. It won't preent the system from shutting down following heat problems, but if it is indeed that, at least you'll know what the problem is (and have some leverage against asking RS for a new fan)
For those of you who have systems freeze up on you during heavy load I would suggest installing MRTG (follow Mouses howto, it's on this forum) and having it monitor your cpu/mb temperature. It won't preent the system from shutting down following heat problems, but if it is indeed that, at least you'll know what the problem is (and have some leverage against asking RS for a new fan)
I was actually trying to install HotSaNIC for this reason.. Servers back up now.. i'll try MRTG thanks madsere
Server crashed again (this time on RS tech support) lol, this is actually kinda humorous.. in a weird kinda way.... MRTG install went well though.... i dunno if i'll ever be able to view it though. lol
Im very thankfull that RS is supporting me with this.. i can not say enough of how much this makes me feel of RS and RS support
Thanks RS!
TT details
Server froze while running memtester, and froze again during another test
compile. Please run the ram through more test passes or check the
cpu/motherboard of this machine.
Im very thankfull that RS is supporting me with this.. i can not say enough of how much this makes me feel of RS and RS support
Thanks RS!
TT details
Server froze while running memtester, and froze again during another test
compile. Please run the ram through more test passes or check the
cpu/motherboard of this machine.
QUOTE
Originally posted by foggy
i'll try MRTG thanks madsere
i'll try MRTG thanks madsere
Ditto Madsere, I will install it as well.
Keep us posted on how things end up on this Froggy. I am not going to start fighting this again until we get the second box and I can move production services over to it.
Once that is done I may just keep this thing and beat it into submission
Rick
QUOTE
Originally posted by Ric
Keep us posted on how things end up on this Froggy. I am not going to start fighting this again until we get the second box and I can move production services over to it.
Rick
Keep us posted on how things end up on this Froggy. I am not going to start fighting this again until we get the second box and I can move production services over to it.
Rick
Here is my sstatus of my server
6/30/02 9:53:19 PM
Will test RAM and install extra case fan.
7/1/02 3:37:25 AM
taking server offline for RAM test
7/1/02 3:43:23 AM
RAM tested bad. testing replacement ram to ensure it's good.
7/1/02 4:07:08 AM
replaced RAM, added case fan. server going back online. if you continue to have issues please let us know
WOW, it made it through a compile and a make.. I think im ok now
Thanks!
im still worried about when there is an actuall load on the server.. i may get a restore though.. it crashed many times the past 2 days and i dont know if that will have a future affect on the server... id rather be safe than sorry and have this resolved before the server is in production....
Does anyone know rackshacks policy for server swaps of troublesome AMD servers?
Darn thing froze up again..
Id be willing to pay some type of fee if i were able to swap this server for another. This server makes me nervous..
who could i contact about this?
Thanks
Darn thing froze up again..
Id be willing to pay some type of fee if i were able to swap this server for another. This server makes me nervous..
who could i contact about this?
Thanks
QUOTE
Originally posted by foggy
Does anyone know rackshacks policy for server swaps of troublesome AMD servers?
Darn thing froze up again..
Id be willing to pay some type of fee if i were able to swap this server for another. This server makes me nervous..
who could i contact about this?
Thanks
Does anyone know rackshacks policy for server swaps of troublesome AMD servers?
Darn thing froze up again..
Id be willing to pay some type of fee if i were able to swap this server for another. This server makes me nervous..
who could i contact about this?
Thanks
back when i was at rs and had a duron
i asked if i could buy a new server and them move the harddrive and i would even move ips but would rather keep same ones and they said no hardware just buy a new server
i was like i will pay for a new server ijust want to keep my data !! they said no so i did back it up but when i did i left.
Im sorry to hear that you had alot of probs too zac.
I just received a responce from my TT and they will gladly swap my server with a celery... So i am a very HAPPY camper
Thanks again RS!
I just received a responce from my TT and they will gladly swap my server with a celery... So i am a very HAPPY camper
Thanks again RS!
QUOTE
Originally posted by foggy
Im sorry to hear that you had alot of probs too zac.
I just received a responce from my TT and they will gladly swap my server with a celery... So i am a very HAPPY camper
Thanks again RS!
Im sorry to hear that you had alot of probs too zac.
I just received a responce from my TT and they will gladly swap my server with a celery... So i am a very HAPPY camper
Thanks again RS!
lol hopefully they are changing the way they treat customers as this WAS'NT an option when i was there.
but overall i'm happier at webreseller.net and dont mind paying 50 more bucks a month for an actual 1U server with good hardware
So who did you contact about it Froggy? I sent an email to sales offering to keep this server and work with RS to get whatever the problem with it is fixed.
All I want is a dependable server for a production box and we would move clients to the new one before tearing into this AMD.
I'll buy another server, I just don't think I should have to pay the setup costs on it. No different than a trade and walk-away from a buggy box, they did not charge you setup fees did they?
Rick
All I want is a dependable server for a production box and we would move clients to the new one before tearing into this AMD.
I'll buy another server, I just don't think I should have to pay the setup costs on it. No different than a trade and walk-away from a buggy box, they did not charge you setup fees did they?
Rick
QUOTE
Originally posted by Ric
So who did you contact about it Froggy? I sent an email to sales offering to keep this server and work with RS to get whatever the problem with it is fixed.
All I want is a dependable server for a production box and we would move clients to the new one before tearing into this AMD.
I'll buy another server, I just don't think I should have to pay the setup costs on it. No different than a trade and walk-away from a buggy box, they did not charge you setup fees did they?
Rick
So who did you contact about it Froggy? I sent an email to sales offering to keep this server and work with RS to get whatever the problem with it is fixed.
All I want is a dependable server for a production box and we would move clients to the new one before tearing into this AMD.
I'll buy another server, I just don't think I should have to pay the setup costs on it. No different than a trade and walk-away from a buggy box, they did not charge you setup fees did they?
Rick
Hi rick, (im foggy by they way, not froggy, lol)
Head surfer has said (i think on WHT or somewhere) that they will replace bad AMD's for the celery's with out any hassle.. So when the server crashed again, all i did was ask in a TT if they would swap it and the responce was "we will gladly swap your AMD server for a celeron 1.3"
I dont have the 60gb hard drive though.. its my old 40 restored, I could have kept all my data, but since that AMD crashed so many times, i wanted everything to be fresh... im ok with that.. at least the server seems more reliable.. but i am getting some segmentation faults (11) i dunno why that is?? im still trying to learn what those are from... im still kinda dumb here... :o
So,anyway... your chances of dumping your AMD for a Celery and still keeping all your data is probably very high
oh, and no.. i didnt have to pay any set up fee
Sorry Foggy.
Well there is no SOP in place for that because sales told me no and suggested I write Headsurfer for an exemption. Which I did... no answer yet but it would be nice to get one before the servers are sold out.
I don't want them to do data swaps for me, I will keep this box and move clients over one at a time. When I get them all moved I can beat this thing into submission if it takes half a dozen fans.
Rick
Well there is no SOP in place for that because sales told me no and suggested I write Headsurfer for an exemption. Which I did... no answer yet but it would be nice to get one before the servers are sold out.
I don't want them to do data swaps for me, I will keep this box and move clients over one at a time. When I get them all moved I can beat this thing into submission if it takes half a dozen fans.
Rick
QUOTE
Originally posted by Ric
Sorry Foggy.
Well there is no SOP in place for that because sales told me no and suggested I write Headsurfer for an exemption. Which I did... no answer yet but it would be nice to get one before the servers are sold out.
I don't want them to do data swaps for me, I will keep this box and move clients over one at a time. When I get them all moved I can beat this thing into submission if it takes half a dozen fans.
Rick
Sorry Foggy.
Well there is no SOP in place for that because sales told me no and suggested I write Headsurfer for an exemption. Which I did... no answer yet but it would be nice to get one before the servers are sold out.
I don't want them to do data swaps for me, I will keep this box and move clients over one at a time. When I get them all moved I can beat this thing into submission if it takes half a dozen fans.
Rick
:confused:
Are you looking for a reduced or no setup fee on a new box and keep your duron untill things are all set.. then move ur clients over to it?? I dont know what they would do in that case, i guess ud have to wait for HS to respond... if ever... with mine, they swapped the server for the celery..no questions asked... and i no longer 'have' the duron...
I thought this would be more attractive to them, no hard drive swaps or anything, I would be doing all the work.
Plus they would be leasing the box while it was being diagnosed and repaired instead of taking it offline. At least I would hope they would repair it and not just dump it on someone as is in the bargain basement.
The compensation they offered you for a buggy box was getting you into a new one without setup costs, that's all I want out of them.
Rick
Plus they would be leasing the box while it was being diagnosed and repaired instead of taking it offline. At least I would hope they would repair it and not just dump it on someone as is in the bargain basement.
The compensation they offered you for a buggy box was getting you into a new one without setup costs, that's all I want out of them.
Rick
Ric I dont want to discourage you but my Raq4i was juat as much a lemon as your AMD box and I was arguing with RS for weeks trying to get a swap - but no way. They did eventually fix the problem with the Raq4i - presumably they replaced the whole thing lock stock and barrel, and eventually the problem was solved, though by then I already had rented an AMD due to the unworkable state of the Raq. (Was reasonably lucky with the AMD though, a dud memory module was only problem so far).
I am sorry to hear that... both for you and for me!
There are other threads on the forum where RS has replaced buggy boxes. I don't think it is unreasonable to expect dependable hardware and it should be clear from looking at the TT history. It is pretty sad that we have to sleep in shifts here so we can keep a 24 hr monitor on this box in case it crashes out of the blue.
This AMD Duron *should* be a better machine than a Celeron, that is why it costs more. It is frustrating because I have never sold or closeted a PC because of hardware problems, I fix them! If we did not have a bunch of clients on it I would fire up a gameserver and drive this thing into the dirt every few hours.
If I have to pay startup charges for a new server, we will pick a different geographical location for it. Do they really charge you a $15.00 reboot fee at webreseller.net Zac? That was the only thing that stopped me from buying a server from them last time I looked.
Rick
There are other threads on the forum where RS has replaced buggy boxes. I don't think it is unreasonable to expect dependable hardware and it should be clear from looking at the TT history. It is pretty sad that we have to sleep in shifts here so we can keep a 24 hr monitor on this box in case it crashes out of the blue.
This AMD Duron *should* be a better machine than a Celeron, that is why it costs more. It is frustrating because I have never sold or closeted a PC because of hardware problems, I fix them! If we did not have a bunch of clients on it I would fire up a gameserver and drive this thing into the dirt every few hours.
If I have to pay startup charges for a new server, we will pick a different geographical location for it. Do they really charge you a $15.00 reboot fee at webreseller.net Zac? That was the only thing that stopped me from buying a server from them last time I looked.
Rick
Well... get this. I had a XP 1900+ with 1GB ram, and it started to corrupt files and such like, caused segment fault with webalizer also. So basically I was a bit knarked off when RS said they were going to swap it for a celery. I paid a massive ($370+) setup fee for this one and now they are going to swap it for something that I could have got a $1 setup on at the time i purchased it. The whole point of getting a more powerful processor with more RAM was blown and I had paid over the odds setup for it.
Now i have managed to get them to put 1GB RAM in the celery. Yes 1GB!!! They do take it which is slightly better than just your bog standard celery, and they allowed me to mount the old hard drive so I could get backups off there.
Now i think i have a network card issue though . Problems are non-stop here.
My first server (AMD duron) was ditched cause of a whole hardware error (they replaced the whole box) and even after that it couldn't handle compiling anything, so i just threw it away, and that was after I had to get it restored because tech support said they would not support me. Which seems ridiculous. The server kept going down. Tech wouldn't support me, so i got it restored. Then it woudld still crash for no reason and they eventually said it was a hardware fault. It could have saved me $70 (priority restore) if they had actually bothered to support me in the first place, considering it wasn't even my fault.
Sometimes I think RS should change hardware vendors.
Anyways, have 2 celerys now (1 with 1GB ram) and they are both pretty darned good. Still wish I hadn't had to pay that massive setup fee or a restore which i shouldn't have had to pay.
James
Now i have managed to get them to put 1GB RAM in the celery. Yes 1GB!!! They do take it
which is slightly better than just your bog standard celery, and they allowed me to mount the old hard drive so I could get backups off there.
Now i think i have a network card issue though
. Problems are non-stop here.
My first server (AMD duron) was ditched cause of a whole hardware error (they replaced the whole box) and even after that it couldn't handle compiling anything, so i just threw it away, and that was after I had to get it restored because tech support said they would not support me. Which seems ridiculous. The server kept going down. Tech wouldn't support me, so i got it restored. Then it woudld still crash for no reason and they eventually said it was a hardware fault. It could have saved me $70 (priority restore) if they had actually bothered to support me in the first place, considering it wasn't even my fault.
Sometimes I think RS should change hardware vendors.
Anyways, have 2 celerys now (1 with 1GB ram
) and they are both pretty darned good. Still wish I hadn't had to pay that massive setup fee or a restore which i shouldn't have had to pay.
James
That is kind of my feelings too, we paid a lot of money for the setup on this machine and if we wanted a bottom end server we would have bought one to begin with.
They did make me an offer for 50% off the setup fee of a new one. That is fair... not good enough to accept but fair.
I can't risk another server here, once we get clients moved they will have our 30 days notice to get this server stabilized while I run an 18 slot gameserver or two. If they can fix it we will keep the thing and if they can't we will dump it.
Rick
They did make me an offer for 50% off the setup fee of a new one. That is fair... not good enough to accept but fair.
I can't risk another server here, once we get clients moved they will have our 30 days notice to get this server stabilized while I run an 18 slot gameserver or two. If they can fix it we will keep the thing and if they can't we will dump it.
Rick
Yeah rick, hope all goes well for you with any new servers you get. I have had 3, and of those 3, only 1 has worked since the start without problems. I have had the other two systems totally replaced (they were AMD systems) and only have 1 of those 2 left. Which is now a celery with 1GB ram.
However I believe I have a network card issue with this one. One problem after the next.
I'm seriously looking into building my own server and getting colo in the UK. Not sure where to go for routers though.
James
However I believe I have a network card issue with this one.
One problem after the next.
I'm seriously looking into building my own server and getting colo in the UK. Not sure where to go for routers though.
James
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.