• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

How Do I - Kill SiteBuilder?

P

PSCGi

Guest
I've tried all the options I can find, including it's docs. YUM fails to
find some 8.6 version stuff and therefore won't remove it. Any other
versions of the command I've tried generate 'SiteBuilder isn't running'.

It started when I accidentally hit SB's # in the Autoinstaller, since
there's no Auto-UN-Installer (makes no sense there isn't) I have no
way of removing it? I have one icon in the CP's that confuses folks
and I'm getting a giant slew of errors in my logs, etc. reporting to me
daily that I don't have SB - duh.

I've got the cap in the S of Sitebuilder correct, tried all 3 doc
methods and nothing's working. Icon in CP's still there and so are
the rest of the resulting problems.

How do you uninstall an autoinstalled piece? Am I missing something?
 
Just the man I was looking for!!!

Seems I've gone to using your stuff in the past because it works, but
here's where I hit a wall. I tried your syntax, but get the exact same
error so I've included the whole action thread below.

Note it says 8.6.0 when I'm on 9.2.1 so I'm confused there. I am
running RH FC8, so that looks proper but it's howling about not finding
stuff on your site?

Mystery is that when I accidentally did this - I used the autoinstaller
located properly and it doesn't give me any indication it's not going to
Plesk info directly as the source. But somewhere - it can't locate the
file it wants to pull this sucker back out!

Here's what happens:

[root@srvded216551647 ~]# yum remove sb-publish
Setting up Remove Process
http://www4.atomicorp.com/channels/plesk/8.6.0/fedora/8/i386/repodata/repomd.xml: [Errno 14] HTTP Error 404: Not Found
Trying other mirror.
Error: Cannot retrieve repository metadata (repomd.xml) for repository: plesk. Please verify its path and try again
[root@srvded216551647 ~]#

Either way with yum - book version or yours I get that. I did try one
other way that generated 'SiteBuilder not installed' and it isn't functional
on my server by any means, nor do I want it there. Just hit a 3 when I
meant a 2 and I'm stuck here. Can't believe the same util won't rip out
what it put in, I've also hit this wall with Qmail...

How? Well I've about had it with this whole Plesk mess and 9.X but I
wanted to stay Qmail. That's been a nightmare and I had clients up my
toto the last few days after the 9.2.1 patch. Finally, I said screw it and
did Postfix BUT I keep getting an asterisk on this same autoinstaller by
Qmail and it keeps trying to stick a 64MB block in all the time that as far
as I've seen - sticks Qmail back in and screws up Postfix, thus hosing
my SMTP again!!!

So how do I uninstall this junk properly and what's up with the missing
'uninstall profile' it seems to be searching for? Am I on your version from
an earlier upgrade that made something work and how Parallels has hosed
this up so bad it's causing that to misfire too?
 
Yep thats a 404 error, it means exactly what that error says.. File Not found. The plesk 8 repo for fedora 8 was EOL'd when Fedora Core 8 was (Last year I think). You can either clean your yum cache (yum clean all) or just remove that repo (rm /etc/yum.repos.d/plesk.repo). Should work just fine for you after that
 
OK!

Got it rolling with your commands, ran the yum clean all but then
for the hell of it did the rm as well. The rm found a file that I would
have thought should've been gone? But for a CYA I did it anyways.

Then running the command below it started just fine as indicated,
but at the end hit a wall on your site (it appears) and that's as far
as I could get? Re-run skips what's already successfully gone with
100% below but runs directly into the error below that.

Any thoughts? Thanks for the (as usual) tremendous help! When
you finally get around to hawking your own OS instead of doing
Parallel's job with Plesk - I wanna know! LOL

[root@srvded216551647 /]# yum remove sb-publish
Setting up Remove Process
updates-newkey 100% |=========================| 2.3 kB 00:00
primary.sqlite.bz2 100% |=========================| 3.7 MB 00:05
fedora 100% |=========================| 2.1 kB 00:00
primary.sqlite.bz2 100% |=========================| 4.9 MB 00:06
updates 100% |=========================| 2.6 kB 00:00
primary.sqlite.bz2 100% |=========================| 2.2 kB 00:00
Could not retrieve mirrorlist http://www.atomicorp.com/mirrorlist/atomic/fedora-8-i386 error was
[Errno 4] IOError: <urlopen error (110, 'Connection timed out')>
Error: Cannot retrieve repository metadata (repomd.xml) for repository: atomic. Please verify its path and try again
 
Still hitting little walls

Still having the issues, though they've trimmed down. Suddenly Logwatch is back, no reason or rhyme but it started working after the last upgrade? Working on configuring it though it appears to have taken a config from what I put in there a looooong time ago and I forgot how helpful that thing is!

The trick is that in the reports, beyond all the pond scum it tracks that I can get rid of with APF, etc. it likes to show me tons of errors in the httpd section about SiteBuilder. That's why I'm trying to get it out totally, beyond the CP icons and menus I'd like gone so as not to confuse people - the biggie is the giant stream of 404 not found errors generated when it tries to boot up what's left of SiteBuilder before that error stops complete removal. It'd make a huge report email chop in half - far easier to concentrate on blocking out the rest of the hackers, script kiddies and idiots trying to hack the box like usual. I mean com'on, some moron hijacked a Sprintlink hotel DSL setup and tried ports scans and hacks that were easily deflected off?

So that's why it's vital to me to un-do what I accidentally did, to make the more important sections so I can flick off the junk that much faster without reading errors I know are garbage.

Thanks!

Dave
 
Here's what I got this morning...

[root@srvded216551647 /]# yum remove sb-publish
Setting up Remove Process
updates-newkey 100% |=========================| 2.3 kB 00:00
fedora 100% |=========================| 2.1 kB 00:00
updates 100% |=========================| 2.6 kB 00:00
Could not retrieve mirrorlist http://www.atomicorp.com/mirrorlist/atomic/fedora-8-i386 error was
[Errno 4] IOError: <urlopen error (110, 'Connection timed out')>
removing mirrorlist with no valid mirrors: //var/cache/yum/atomic/mirrorlist.txt
Error: Cannot retrieve repository metadata (repomd.xml) for repository: atomic. Please verify its path and try again
[root@srvded216551647 /]#

I'll attempt to scope it out myself, but what IP am I shooting for and what port? Maybe APF's cutting this flow off, not seeing it in the logs but stranger things have happened. I can at least unblock a specific path and maybe that'll open things up. I thought you were domestic, but if you're in RIPE or APNIC territory I can almost guarantee there's a block doing something here. It gets to that point, then that's it.

I got the usual slew of overnight reports and the Logwatch report of problems based on SiteBuilder, so any way I can rip this sucker out the better! LOL

Thanks!
 
Ugh, I hate this...

Man! I hate doing this because somehow it gets taken personally, even though I've been there and done that myself... I ran several traceroutes, pings and whois's to not only atomicturtle.com but also websitewelcome.com, the NS's used for it and what I'm supposed to be snagging.

I hit major time out walls. Over and over. Thought it was my end, so I ran several tests from various points inside that HSP. Same thing. Thought it was still them, wouldn't be a shocker, so I went and ran tests out the wazoo from various national and international points. Ugh.

EVERY time I get timed out reports along the way from the starter point down through various backbones and then the bombs when and continuing from The Planet/EV1 down to it's Dallas DC and your box(es). NS's, domain and all - massive bombs. I can easily see why I can't get this thing, I'd have to run it over and over for hours before I'd get lucky and hit my way in to get that file. That's way over the limit.

I know it's not your fault, I believe I found you originally on EV1 back when we could enjoy the place - I'm not there anymore. ThePlanet destroyed everything EV1 built and tanked it with their own junk. Stuff like this was routine and they refused to even acknowledge stuff when I provided reports and alternative testing reports/locations they could do it. They saw it was fine 'to them' inside their own little world, therefore we were all nuts.

So how do I work around this? How can I snag this without hitting those IP's/NS's and get it gone? And now I'm really worried because I have your package service installed over Plesk's and what else am I going to have issues with as long as you stay there? Not trying to kill your business as you don't really 'charge' per se and I know over the last 5-10 years I've watched and worked with you I trust your stuff far over even Parallel's junk, but I can't get stranded even unintentionally.

Can I wget files somewhere to help this? Will your software find it if it's on my box and then complete this? I'm trying to find an easy solution, my ultimate suggestion is do what I did and run like hell from ThePlanet before it's too late but that's just me. Anything I can do to resolve this meanwhile I'm for it!

Dave
 
Miracle!

OK, I pounded and pounded it, running traceroutes and pings between attempts to hopefully get the one magic hole for what turned out to be a 952 byte file? Sheesh! LOL

Finally - got it, erased everything and hopefully that's that!

Did take quite a few attempts and such, I also ran those alternates from various points around the US and then overseas, they all seem to fly right up to Level 3 - then once hitting ThePlanet and into it's Dallas DC, we hit those errors constantly. And I was told this same thing like 2 years ago before I got the hell outta there, so I know it's not just me on this. I was read the riot act from clients and more about such problems, intermittent website access, FTP's dropping mid-stream and worse. I stayed till the final straw, I talked way up the food chain to the 'final' crew that remained at ThePlanet and was insulted so bad we walked away totally and took the heat of the switching last-second like that. But, afterwards it's been peaceful and far more enjoyable!

Just an FYI!
 
Well now, ain't that strange!

Wonder why BOTH of those URL's come back as ThePlanet and the Dallas DC?

One I could see, but both were routed that way from all points, from the HSP
testing point through to Level3 and then through to TP and it's Dallas DC, from
the looks of the coding on that it's old EV1 server DC they bought up. Didn't I
originally see you a few years back on EV1's forums along with here? That's why
it didn't seem strange to me.

You got any DNS's roaming around you don't know about or that didn't update?

Side Note: Removed it all, but didn't take cronjobs out and I'm going insane
trying to swat flies for 404's and missing directories. I renamed the SB dir that
was left for some reason and that's tripped the cron's - cept they're not easily
listed in the crobjobs I can find in /etc to kill them off since they do nothing!

Maddening...

I'll rescope the domains again, show you the results I get if they show what
they did over the weekend!
 
No, nothing like that. What do they resolve to? Maybe someone at theplanet has set up a transparent proxy or mirror? I know internally theplanet uses the atomic repo pretty heavily.
 
I have NO idea.... I ran it over and over again when it was fixed and now it all comes up as you mentioned. Not sure at all why I'd even hit a transparent layer like that, much less from multiple locations. One I could see and investigate but we're talking like 10 so I wasn't talking outta my (#*$!! LOL

And ThePlanet... ROFL, well that doesn't surprise me at the least not only saying volumes about Parallels but then also even more volumes about TPL to boot.

So far it's all worked, problem is I've got scattered leftovers and references that didn't get removed and now I get slews of cronjob errors I can't find the commands for, cron overnight errors, etc. I'm all over the /etc directory but I find no reference at all? Presumably it's hiding somewhere else, even though these seem to have some effect when changed, but I can't for the life of me find the SB references in cron's I know of to make those jobs stop and give me some peace and trim my logwatch down to normal again.

Dave
 
Back
Top