Why I left pair.com after 12+ years and why that makes me sad

// November 19th, 2013 // Rambling

UPDATE: I’ve moved most of my former Pair-hosted sites to Media Temple and I’m also managing a client’s hosting there as well. I’m pretty happy with MT overall for managed-hosting needs. Agghosting and its clients is running on its own server over at Linode and I’m super happy with this [unmanaged] machine.

I recently had to abandon ship from my preferred hosting company of some 12+ years, Pair Networks. Seriously. I once loved them and recommended them left and right. The falling apart started, I now recognize, once I moved from my old and no longer offered plan to their new shared hosting. It was fine for a while. I didn’t host everything there — this blog, for instance, isn’t — but a handful of sites for which speed mattered to me (and some I moved when space became an issue, prior to the move to the ‘new’ system) were there.

Now they’re at Media Temple and they’re fast and reliable again. And the price is the same. I just feel sad that Pair let me down. Like an old friend who suddenly snubs you. I was planning to move Ruining It for Everyone there, but the performance failures made that untenable. It’s still at Dreamhost which was meant to be temporary (and is still more reliable and faster than my Pair account had become). I’ll move it to Media Temple soon, too.

But back to my Pair problem… Sorry. This is long and I didn’t originally intend to spew all this, but it’s happening now so just hang in there… Over the course of 4 days I buckled down and reported the problems I’d been having. (It was to the point that I had to start silencing my phone at night since I’d get a slew of ‘site XXX is down’ texts which would wake me – which really wasn’t a good idea for me to do considering I run a computing organization and should be woken up should something go wrong there.) Given my past experience with them, I felt confident that someone would sort out the issue and everyone would be happy. Not the case.

Basically, I heard ‘it’s OK now’ and ‘the issue was resolved and load is going down’ and ‘sometimes shared hosting gets slow’ and ‘we monitor our servers 24/7, you don’t need to send process and load info’ and… basically a whole fucking lot of nothing.

It goes like this:


On Apr 25, 2013, at 10:11 PM, Margaret McFee wrote:
My site(s) have been going down with fair regularity. It’s happening
now so I hopped on the server and this is what I see:

12:30:19 up 13 days, 18:00, 3 users, load average: 535.59,
307.70, 138.51

Machines is Linux www2.g1.pair.com 2.6.32-46-server #107-Ubuntu SMP
Fri Mar 22 20:30:21 UTC 2013 x86_64 GNU/Linux

Could you please look into this?

Thanks.


Apr 25, 2013, at 10:08 PM
Hello,
Thank you for contacting pair Networks. We apologize for the inconvenience
you have had with your server. We were aware of the issue and have been
trouble shooting the issue all day. The server is now up and running and
the load averages are dropping as was expected. Again we apologize for this
inconvenience.

Best regards,
XXXXXXXX
pair Networks, Inc.
support@pair.com


** Load is NOT down, despite this assurance at 10pm **

On Apr 26, 2013, at 12:58 AM, Margaret McFee wrote:
If it helps, it’s currently at really high load
00:54:05 up 14 days, 6:24, 4 users, load average: 574.34, 461.68, 261.97

Here’s what’s running at this moment:

% ps auxwww
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND

(long ps out put follows)


** 1am and now I’m feeling brushed off **

On Apr 26, 2013, at 1:07 AM, Margaret McFee wrote:
Another ps snapshot a short time later. It’s very long. Sorry.

mmcfee@www2% uptime
01:02:25 up 14 days, 6:32, 4 users, load average: 331.02, 375.95, 292.79
mmcfee@www2% ps auxwww
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND

(very long ps output – I hoped they’d look and see “Ah, it is not resolved after all, is it.”)


On Apr 26, 2013, at 2:58 AM, pair Networks Support wrote:

Hello,

Our systems administrators are currently aware of and working to stabilize
your hosting server. I apologize for the inconvenience caused by these
problems, but we are working to have the situation rectified as soon as
possible.

Also, we log the full processlist automatically every few minutes, so there
is no need to submit this output to us in a support ticket.

Best regards,
ZZZZZZ
pair Networks, Inc.
support@pair.com


** A day goes by, same behavior. Phone and email abuzz with ‘site down’ messages throughout the day and night. **

On Apr 28, 2013, at 3:20 AM, Margaret McFee wrote:
Hi, guys.

Still a regular occurrence. No change over the last 3 days that I can see. On Friday it was down/unresponsive 5 or 6 times through the day (that I’m aware of) and slow at other times.

This is getting to be a real concern and is keeping me from moving a site. Well, I did move it, actually. But then had to switch the DNS right back because I’d basically taken a working site and crippled it… I really don’t want to have to consider putting it on another host. Please let’s not let it come to that… I’ve been with Pair for 12+ years and this is the first timeI’ve ever considered jumping ship.

Sites are currently inaccessible again and load is through the roof again.
02:57:04 up 16 days, 8:27, 8 users, load average: 469.62, 285.15, 187.29

Thanks,
Maggie


On Apr 28, 2013, at 4:57 AM, pair Networks Support wrote:
Hello,

Load is currently quite low on the server and your sites seem responsive at
this time.

If possible, please contact us at urgent@pair.com the next time you notice
the problem occurring. We can then give the matter our immediate attention
and hopefully narrow down the cause.

If you have any other questions feel free to ask.

Thanks
YYYYYY
pair Networks, Inc.
support@pair.com


** OK. Next time I’ll send email to urgent@pair.com instead of support@pair.com as directed. We’ll see how that goes. **

On Apr 28, 2013, at 12:53 PM, Margaret McFee wrote:

OK. Site is unresponsive right now.

12:51:54 up 9:23, 3 users, load average: 193.66, 125.06, 66.13

PS – I was previously stated you were aware of this and actively working on it. This reply implies otherwise (ie – “it’s OK now so there’s no problem”).

>
> [quoted email from YYYYYY saying to contact urgent@pair.com]
>


On Apr 28, 2013, at 10:31 PM, pair Networks Support wrote:
Hello,

It appears that your hosting server was under high-load at the time that
you submitted this ticket, however, currently the server is running
normally and there is no indication of a problem.

Our servers are monitored 24 hours a day and the issue appears to have been
resolved promptly. You may contact our urgent technician at
urgent@pair.com for downtime related issues so that the problem can be
investigated as it is occurring.

Sincerely,
AAAAAA
pair Networks, Inc.
support@pair.com


** The benefit of doubt has been used up, “urgent” took 9.5 hours to respond, their response (9.5 hours later) was ‘looks fine now’, and now I’m ticked off. **

On Apr 28, 2013, at 11:31 PM, Margaret McFee wrote:

So, just to be clear, you’re saying the issue I’ve been having over these several days is resolved?

That the repeated excessive loads (in the 300s, 400s, 500s) I’ve been reporting — the prior reply telling me to contact urgent@ instead of support@ should it happen again, which I did — for the past 3 days is no longer an issue, was observed by your 24/7 monitoring (not requiring me to report it) and action was taken which resolved the issue today in the 9.5 hours between my ticket and your reply below?

Or are you saying that it was fine when you looked and therefore there’s no problem as far as Pair is concerned?


Hello,

There may be periods of high-load on a shared-hosting environment, as there
are many websites being served. These instances of high-load are seperate
issues most often, and a site that is causing a chronic problem will be
addressed by our administrators. By writing to our urgent technician you
may ensure that they are looking into any current problem that you are
experiencing.

As the server is currently functioning normally, there does not appear to
be an issue that I can investigate at this time. Is there a website that
you are currently having trouble accessing?

Sincerely,
AAAAAA
pair Networks, Inc.
support@pair.com


** May 12th I move the last domain off Pair. May 13 after work I start deleting the domains from my account. Moments after the last one, Billing sends an email. My, what great response time billing has! **

On May 13, 2013, at 5:55 PM, pair Networks Billing wrote:

Hello Margaret,

We were reviewing your billing and noticed you cancelled your domains. This does not cancel the billing account. Do you want to have the entire account cancelled?

Thank you,
BBBBBBB
billing@pair.com
pair Networks, Inc.


I reply yes, yadda yadda.


So tonight when I got home (some time after 1:30 or so) I notice an email asking me to take a feedback survey. And here I am at 3am venting my spleen. You’re welcome to read my response.

Customer Service Survey

Please take a minute to answer the following questions regarding your recently resolved customer service or support issue.

Was this issue resolved capably and correctly?
Definitely Not

Are you satisfied that this issue was resolved promptly?
Definitely Not

Were you given friendly and courteous assistance by our representative(s)?
Not Certain

Well, I left and closed my accounts, if that’s an indicator. I did so because no one seemed to care that my sites were constantly slow and often unresponsive even when I presented evidence of systemic unresponsiveness and _extremely_ high load averages. So after more than 12 years of hosting with and recommending Pair, I’m now at Media Temple paying the same but for far more speed. Add to this that (mt) noticed me saying I was fed up and moving to (mt) and even offered to help (I’d already paid, they didn’t need to lure me), whereas Pair didn’t notice until I deleted my last domain. Oh, but the billing people responded within minutes, if not seconds… And no, I don’t need a management response. I’m no longer your customer nor your cheerleader so I, quite frankly, could not care less what happens from here on out. Good luck. I do actually do hope I’m an edge case (well, me and everyone on that miserable cluster) and not an indicator of Pair’s descent into mediocrity.

See B4H6V2P for timeline.

PS – 3rd question answered “not certain” because your techs were courteous, despite dismissing me with shrugging ‘it’s fine now’ answers. I never played the ‘Come on, I deal with high-demand computing all day’ card with them because I’d hate it if it happened to me. But I do. I gave them, initially, the benefit of doubt, especially as past experience was that Pair has/had quality ops people. I’m not so sure as I once was. So in the end I did ask some pointed questions to elicit a response. All I heard was “It’s fine now”. So if that level of performance is “fine”, I need to go elsewhere.

Comments are closed.

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