Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


BuyShared.net
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

BuyShared.net

sureiamsureiam Member
edited January 2018 in Outages

Soooo... Anyone else noticing about 1 hour total down time a week pretty consistently on the shared platform by @Francisco ?

Love the features and price but starting to feel it would be wiser to move my static websites to a cheap VPS. Dynamic sites on buyshared are a lost cause way to slow. But cached dynamic sites to static sites load incredibly fast (as to be expected). I wouldn't even mind so much if it was 1 min here or there but they are like 8-10+ min blocks during peak daytime hours. These also are low traffic sites. Only about 10 or so unique visitors a day max. So it's not like I'm a trouble maker here.

This has been discussed with them with little consistent resolution. I know the price is cheap but so is a good VPS. Am I expecting too much here? Have others also noticed this?

I'm also a little grumpy because today there has been about 20 mins down time without warning.

«13456711

Comments

  • Did you confirm that there was actual downtime? IIRC the apache (or nginx) process gets soft-restarted every time someone on the server adds/removes a domain, which will cause uptimerobot and the likes to generate a downtime report, and will recheck quite a bit later. Meanwhile, the downtime should barely be noticeable, if at all.

    tl;dr: did you try to load the site when your monitor said it was down?

  • sureiamsureiam Member
    edited January 2018

    @teamacc said:
    Did you confirm that there was actual downtime? IIRC the apache (or nginx) process gets soft-restarted every time someone on the server adds/removes a domain, which will cause uptimerobot and the likes to generate a downtime report, and will recheck quite a bit later. Meanwhile, the downtime should barely be noticeable, if at all.

    tl;dr: did you try to load the site when your monitor said it was down?

    3 geo located monitors requiring 2 to fail and about 8 mins of down time checked each minute to even trigger an alert to be sent..

    This is legitimate down time. But I can understand your response as many probably down monitor their sites this carefully.

    As noted i wouldn't mind if they were 1 min occasional burps which is why the trigger is 8 mins with 2/3 locations failing before an alert is sent. Note there have also been some longer ones like 30-40 mins but those are rare and @Francisco is a good guy and explains the steps taken to resolve that issue. So I've been pretty quiet about it. But at some point i feel the community here should discuss this. Its a cheap service though so Im not outraged just thinking if it's time i move my static pages to a VPS or if these outages can be resolved.

  • Yes, can confirm. Hosting a small site with them and I had WordPress Jetpack monitor set up. Got 4 alerts today with 8 to 10 minutes of down time.

  • WHTWHT Member

    This have been happend in the past and the issue was on litespeed as I can remember.

  • Its not really a secret Buyshared is not quite the quality that some say it is

    Thanked by 3Xei ucxo AMXRT
  • I changed providers and went to Ramnode. Prices are not comparable though. I have to pay more now for that uptime.

    Thanked by 3Ympker lazyt AMXRT
  • Are you by any chance on lu02? Ask franny to move to las vegas, it's solid.

  • For the price its ok

  • You are not alone. When I was with BuyShared, I also got many small downtime here and there. There was a consistent problem where the server load would sky-rocket during the daytime so maybe that contributed to the instability? There was also a problem with phpMyAdmin that it just took ages to load. It should not take me 5 minutes to see what my tables look like in phpMyAdmin. This was just my experience when I had BuyShared in their US location. I didn't create any fuss about it with them since the price is dirt cheap so I just peacefully moved to a different host.

  • When I was with BS, it happened a few times. But, it was cheap and I did not mind. Later moved out.

  • sureiamsureiam Member
    edited January 2018

    @jetchirag said:
    Are you by any chance on lu02? Ask franny to move to las vegas, it's solid.

    No this is Vegas.

    @WHT said:
    This have been happend in the past and the issue was on litespeed as I can remember.

    Yes, this is the reason I've gotten every time also.

    @Amitz said:
    I changed providers and went to Ramnode. Prices are not comparable though. I have to pay more now for that uptime.

    I'm starting to feel the same. Its a shame we can't get better uptime though. I don't expect performance to be high (hence only static pages there) but uptime could be higher.

    @Mozilla said:
    You are not alone. When I was with BuyShared, I also got many small downtime here and there. There was a consistent problem where the server load would sky-rocket during the daytime so maybe that contributed to the instability? There was also a problem with phpMyAdmin that it just took ages to load. It should not take me 5 minutes to see what my tables look like in phpMyAdmin. This was just my experience when I had BuyShared in their US location. I didn't create any fuss about it with them since the price is dirt cheap so I just peacefully moved to a different host.

    Absolutely the load on the servers is generally very high as the users tend to run too many dynamic pages on a shared host (you can search through IP bgp to easily see the other users on the node). But if you run a cached dynamic page as static the performance is steller. Litespeed though is known for its static performance so not too surprising there.

    Of course you get what you pay for but I can't imagine @Francisco would tarnish the buyshared and buyvm reputation with sub par uptime. There's a reason I'm not really looking at buyvm now to migrate the Sites over to....

    Troubling users and resellers should be taken care of in my mind. If a $25 a year reseller has 1000 sites on there or a $10 a year user has 500 sites buckling the node I would expect some sort of response to be taken.

    Perhaps @Francisco can consider putting some reasonable restrictions on his services to retain the responsible users that only keep static sites and shared resource to a minimum.

  • Same here, but I didn't mind, I had a $2/month reseller, and 5~10 mins downtime a day is not a problem to me. But later some of my clients were monitoring their site with uptimerobot and started complaining about the 5 mins downtime, so I moved out to a VPS..

  • AmitzAmitz Member
    edited January 2018

    sureiam said: I'm starting to feel the same. Its a shame we can't get better uptime though. I don't expect performance to be high (hence only static pages there) but uptime could be higher.

    I really do not want to talk bad about @Francisco and his services - It seems that many people are more than happy with what he provides. But BuyVM has never been the right fit for me, neither with shared hosting, nor with VMs. That happens in life.

    Thanked by 2shovenose netomx
  • @Amitz said:
    I changed providers and went to Ramnode. Prices are not comparable though. I have to pay more now for that uptime.

    Also with Ramnode. No issues at all so far. Just the announced 5 minute reboot due to the Intel incident.

  • @Amitz said:

    sureiam said: I'm starting to feel the same. Its a shame we can't get better uptime though. I don't expect performance to be high (hence only static pages there) but uptime could be higher.

    I really do not want to talk bad about @Francisco and his services - It seems that many people are more than happy with what he provides. But BuyVM has never been the right fit for me, neither with shared hosting, nor with VMs. That happens in life.

    I dont wish to speak negatively of him or his services either. For what you get the pricing is very reasonable. However I believe better uptime can be achieved and opening this up for discussion.

  • I had the same troubles in luxemburg.. 2 - 3 downtimes about 10 minutes a day . I opened a ticket and got friendly support from Francisco himself
    He said something about a failed backup which caused the downtimes

    He solved it, now I have one 10 minute downtime once a month

    But this happens because of peering problems in luxemburg which is a known problem

  • raindog308raindog308 Administrator, Veteran

    Hybrid said: Same here, but I didn't mind, I had a $2/month reseller, and 5~10 mins downtime a day is not a problem to me.

    That's where I am, too. It's as good as most shared hosting. If you want super good, then pay more. If you can live with good enough, BuyShared fits.

  • Shared hosting is for those sites you want to shove on a reputable operator and forget about it for 18 months at a time.

    BuyVm/Buyshared fits the bill for me at @20/yr.

  • FranciscoFrancisco Top Host, Host Rep, Veteran
    edited January 2018

    fburger said: But this happens because of peering problems in luxemburg which is a known problem

    Once we moved things away from Root's network that finally calmed down. They have an autonull of sorts that would flap our port whenever they saw anything resembling a flood. Their techs deny it but to be fair their techs also didn't know they had BGP communities until I talked to iptransit.

    This morning we had 2 burps in LV. I didn't see a maintenance email get sent out over the holidays that one of our upstreams was doing router maintenance so that was entirely on me dropping the ball.

    We still get the Litespeed burps that last a minute or so, but by no means anything going on an hour though. Litespeed has a 'watcher' script that does help when cPanel decides to hard kill Litespeed, but that doesn't seem to help with the 1 minute restart timeouts.

    If you're getting hour long black outs I'd for sure ticket to see if you're getting nullrouted. I'd recommend you ticket and get my personal email to put on whatever monitoring script you use. I got a few users that do that, and that's how we figured out some issues for them.

    Thanks!

    Francisco

  • Francisco said: I'd recommend you ticket and get my personal email to put on whatever monitoring script you use.

    RIP your inbox, as they say on Reddit.

    A dumb idea: is there a way to make sure that Liteserver is setting SO_REUSEADDR ? Leaving it out could interfere with the restart for a minute or so.

  • @Francisco said:
    Once we moved things away from Root's network that finally calmed down. They have an autonull of sorts that would flap our port whenever they saw anything resembling a flood. Their techs deny it but to be fair their techs also didn't know they had BGP communities until I talked to iptransit.

    Where have you moved your servers to? I've left BuyShared when my account was on lu04 (IIRC one of the CentOS 7 servers there), if you're now on a more stable network I'll perhaps give another try.

  • FranciscoFrancisco Top Host, Host Rep, Veteran
    edited January 2018

    quicksilver03 said: Where have you moved your servers to? I've left BuyShared when my account was on lu04 (IIRC one of the CentOS 7 servers there), if you're now on a more stable network I'll perhaps give another try.

    We didn't move them anywhere, we're still in Root :) I simply put in another bandwidth provider. I can't balance the bandwidth between them and the other provider yet since the router we sent there doesn't have enough 10gig ports.

    I got a Juniper MX80 sitting in Vegas pending shipment there but I have to get them to move us to another cabinet since they made a huge mess when they put my slice nodes in. I'm likely postponing this until I can send a slab node along for the ride as well as a handful of slice nodes, that way I can button up the location. Past that, I don't think I'm going to send anything more there, it's frustrating dealing with it.

    willie said: A dumb idea: is there a way to make sure that Liteserver is setting SO_REUSEADDR ? Leaving it out could interfere with the restart for a minute or so.

    I've watched it happen before and I think it's just cPanel gets impatient. It's likely a combo of being a lot of domains and it taking a few to read the httpd.conf. Litespeed takes a few to get everything brought back up since it's parsing through the file. Litespeed itself forks off another process when it's restarting things to get around this (so its gotta be using REUSEADDR), so I'm thinking it's just cPanel getting grumpy.

    I was thinking it was not getting enough CPU time to complete things but we bumped lv-shared03 to 40 CPU threads (dual 2660v2's) and that node sits like 50%+ idle and it still does it.

    I've ticketed cPanel about it and they usually figure it out and merge a patch (or just adjust timeouts maybe) but then it breaks a couple weeks later.

    Francisco

  • @Francisco why is it raining teeth on BuyVM's website?

    Thanked by 1Aidan
  • It's Caries Day in Canada.

    Thanked by 1Aidan
  • FranciscoFrancisco Top Host, Host Rep, Veteran
    edited January 2018

    @Harzem said:
    @Francisco why is it raining teeth on BuyVM's website?

    Those are horse heads :P I should remove that for now. Next year i'll make the images a bit bigger me thinks. You can see the full sized one at https://manage.buyvm.net/

    I'm going to check my logs to see which node has the most 1 minute blips so far and upgrade it to a Litespeed 2CPU license. While I don't see the Litespeed process riding top very much (mostly just when it's restarting), maybe it needs it.

    Francisco

  • HarzemHarzem Member
    edited January 2018

    Francisco said: Those are horse heads :P

    HAHA I never thought of that. They looked like teeth to me :) Maybe put a "powered by stallion" or something with the large logo somewhere, so when people see the small logo, they can understand what it is.

    You could even make that horsehead your logo, left of "BUYVM" at the top. Wouldn't necessarily make perfect sense but wouldn't look so bad either.

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    Harzem said: They looked like teeth to me :)

    You aren't the first! As I said, i'll make them a bit bigger next year to make it more obvious.

    I tried that out but I didn't like it. I've grown to like my simple 'text with a gradient' on it.

    Francisco

  • I'm on lv-shared01, and I seem to see some sort of downtime everyday from Uptime Robot and occasionally from JetPack on my WordPress install. It doesn't really bother me though. I am guessing that Uptime Robot errors are false positives.

    The site always seems to be there when I visit, and they aren't of any importance to me.

  • teeth. You are high.

    @Harzem said:

    Francisco said: Those are horse heads :P

    HAHA I never thought of that. They looked like teeth to me :) Maybe put a "powered by stallion" or something with the large logo somewhere, so when people see the small logo, they can understand what it is.

    You could even make that horsehead your logo, left of "BUYVM" at the top. Wouldn't necessarily make perfect sense but wouldn't look so bad either.

  • @Francisco said:
    I'm likely postponing this until I can send a slab node along for the ride as well

    How goes the slabbing process anyway? Looking forward to giving you another try in LV when it comes around.

Sign In or Register to comment.