Ever since one my sites was moved from WIN21 to WIN33 I’ve had terrible uptime. Today it went down for at least an hour and a half. This site was running with much less difficulty on WIN21. My understanding was that WIN333 was an upgrade but so far it seems like it cannot handle the load and the application pools are constantly hanging. I’ve been told that my site is in a dedicated app pool but it should recycle sooner than an hour and a half.
Anyone else having issues on WIN33? Is Jodo aware of an ongoing issue? Do I need to move this web site to a more stable option?
I am not aware of anything near an hour and a half today, but we found the problem on it that made the issue today.
Win21 was down a LOT more than win33 ever has been even with the inital hardware troubles(since moved off that hardware)
in fact win21 is basically ready to die, the hard drives keep going offline all the time, breaking raid, no logs work on it, etc. The server is still up but only suspended accounts are on it, since they can’t be moved, we will be working this out shortly.
There is absolutely no overload to win33, it is much more machine than win21 was, if you are having ongoing problems with the pool Manish or myself can investigate it but you need to send email to helpline really, as that is the proper channel.
(it is helpline @ the domain here minus support part)
The issue today was actually log files and I am working on a resolution for it. (I only know of about 15 min max of iis being slow then a reboot to resolve it, I was on the server 1 hour before with no issues so know it wasn’t down that entire time) The monitoring system told us of problems when they happened making iis non responsive, but again we cannot monitor every single domain/pool on the server.
We have setup 2 monitors on win33, at two different pools, but there are more pools that than, and I see two client custom pools as well.
Thanks for the reply. I understand that WIN33 is a better server than WIN21 but I just haven’t had a good experience with it to this point. The initial switch over caused some major downtime and it does not seem to recover quickly when it has issues. Win21 used to go down but it always seemed to come back online without much effort on my end.
I set up a monitoring service, at my clients request, for this site in question in August of last year. It pings and does a content check. The reporting feature shows that the site has been down for 1 day 1 hr and 15 minutes in the last 3 months and 1 day 7 hrs and 41 minutes in the last 6 months. With 6.5 hours difference and the majority of the downtime occurring after 2/6/2010 WIN33 hasn’t been better for my client’s site.
Additionally, I was checking the site the whole time it was down today and it definitely would not load for at least an hour and a half. I also received multiple error alert emails during this time from associated web sites that share content. I don’t mind late night outage but daytime outages cause customer service issues for me.
I generally expect, from my 12 years of reselling hosting, that shared/reseller hosting has a 99% uptime, with an approximate total possible downtime of 3.5 days a year. WIN33 or at lest my client’s site has been down, unresponsive, hung, etc… for at least one day in the last month and a half.
Some of this downtime could be caused by the fact that the site needed to be put in a dedicated application pool, as it was on WIN21, but shouldn’t the shared app pool recycle every 20-30 minutes? I’ve had to put in 6-7 new support tickets for this site being unavailable since February and I generally only submit a ticket if the site is down for mor than half an hour.
I also have gotten pretty short replies to my tickets so I’m not clear about the longer outages. Usually something about reseting or moving app pools. I asked today whether I could copy my monitoring alert emails to a Jodo email address to help facilitate a resolution but I did not get a reply. Is this something I can do to help recycle the app pools more quickly when the site goes offline?
I’m not trying to complain, just trying to get figure out what is going on and make sure it is not a bigger issue that is going to cause more problems. I’ll try to remember to use the email address you provided but I usually go straight for the helpdesk site support tickets. I would appreciate anything you can do to alleviate this issues so I don’t have to worry about it. Thanks!
helpline isn’t a replacement to the ticket desk it is just a supplement for management review.
No we really can’t take a copy of every email from monitoring, it will make big problems in the ticket desk and result in blocks on the ticket system, so please don’t go that route (the auto replies and such tend to get into wars)
Use the chat function and the ticket desk to let us know, and we’ll work to get to the bottom of this for you.
also just as a FYI, no recycles don’t happen every 20 minutes or 30 minutes, yours was actually on about 2 hours to recycle. I can set it lower, but sessions and such on sites will start going kaput if set for less.
Ok, I’ll continue to use the chat and helpdesk to report issues if they do not fix themselves withint 20-30 minutes. My concern is that I will be out of the office and without email access, then the site will hang, I wont get the email alerts, and without follow up to Jodo it could be down for 2 hours or until I follow up with a a support ticket. Any way to reduce this concern? Or, is there more monitoring going on to proactively fix these issues?
Regarding the recycles, I was thinking about sessions that have a default timeout of 20 minutes. It looks like, according to a search on Google, that the default recycle time is 29 hours.
Thanks again for the replies Stephen. Can you address my other questions shown below?
My concern is that I will be out of the office and without email access, then the site will hang, I wont get the email alerts, and without follow up to Jodo it could be down for 2 hours or until I follow up with a a support ticket. Any way to reduce this concern? Or, is there more monitoring going on to proactively fix these issues?
The site has been pretty stable since the 16th so a 7 day monitor may not do much. I also have monitoring on this site so I will get alerts if there is a problem. My concern is that the site will go down in the future and If I’m not here to submit a support ticket, it will stay down for too long because a support ticket is required to resolve the issue.
Once we get our new monitoring server up and going we will be able to put a monitor on it for longer. I have started working on the new monitoring center this week, but not finished it.