HelioHost

Heliohost is currently experiencing a DNS issue on our main domain http://heliohost.org/ and http://johnny.heliohost.org/ Attempting to access these domains by name will most likely result in a server not found error for most people. We are aware of this issue and are working to resolve it. All user DNS data seems to be intact, and our forums http://helionet.org/ are working fine. If your account is located on johnny you can reach your cPanel by using this link http://64.62.211.131:2082/ since the server is online and fully functional. We apologize for any inconvenience this unexpected downtime has caused you.

If you're one of the 2286 co.cc domains hosted on Heliohost service you may have noticed that your website hasn't been working the last few days. This is NOT because Heliohost is offline, but rather the http://co.cc/ service is offline. Heliohost has no control over what domain registrars do so unfortunately we can't get your domain working again. If you've been affected by this co.cc downtime we recommend using parked domains and addon domains to get around the problem to make your website visible again. You can also use this tool http://www.heliohost.org/home/support/scripts/domain to change your main domain. Hopefully co.cc is back online soon, but in the meantime if you're having any trouble settings up a new domain let us know and we can help.

Since the load on Johnny seems to be settling down a little, Java has been upgraded and enabled once again to version 7.0.32. MySQL on Johnny has been upgraded from 5.5.22 to 5.5.27, and cPanel has been upgraded from 11.32.2 to 11.32.5. During the downtime and the subsequent high load following Johnny's crash there ended up being about a thousand backlogged accounts in the creation queue. We're happy to announce that all of those account creations have now been processed and the accounts should now be active. If you've been waiting patiently for a week or more for your Johnny account to be created you can now log in at http://heliohost.org/ It's been stated several times already, but inactivity suspensions and automatic deletions have been postponed on Johnny due to the inability of users to log into their accounts since October 3rd. It has now been a full week since Johnny came back online, but since this last week has only had about 50% uptime due to frequent spikes of high load we are going to extend the no-inactivity period until November 1st. Be sure to log into your account before this date if you don't want to end up suspended for inactivity. This should probably go without saying, but the offer for backups has expired since the server is now online. If you wish to create a backup of your Johnny account you can use this link http://johnny.heliohost.org:2082/frontend/x3/backup/index.html to create the backup yourself. We recommend making partial backups not full backups, and saving the files to a safe location often to prevent data loss. Since the server is back online you no longer need an administrator to manually delete your account for you. If you wish to delete your account simply fill out this form http://www.heliohost.org/home/support/scripts/delete and your account will be deleted. Account deletions are highly dependent on server load so if the load is high it will take a while to be deleted or if the load is low your account might be deleted instantly. It just depends on the specific circumstances at the time.

Johnny is now back online and everything seems to be functioning. Unfortunately the postgresql backup and restoration failed so all data that was stored in postgresql has been lost. We apologize to anyone who has lost data in this way. If you have a backup of your postgresql databases you can import them back into your site. Let us know if you see anything else behaving strangely and an admin can take a look at it. Thus ends the longest downtime that Johnny has experienced since he first came online over two years ago. Welcome back Johnny; we all missed you quite a bit, and we're glad you're back.

We've got some good news and some bad news to announce with regards to the server, Johnny. The good news is it isn't a hardware problem. All of the physical components of the server appear to be in tiptop shape and are performing perfectly. The CPUs are all still working, the hard drive didn't crash and all user data appears to be intact, the memory hasn't gone bad and passes all tests we have run on it, the BIOS is still operating, and everything else seems to be operating optimally too. The bad news is the operating system appears to be corrupt and the system is unbootable. This means that when the server is started it crashes during the boot process, and no services such as http, ftp, or even ssh are ever running to give the admins any kind of remote access. The good news is that we have managed to boot up Johnny from an operating system located on a virtual boot disk with limited networking capabilities. The original filesystem on Johnny has been mounted and is completely accessible to us. We can confirm that all user data is intact, and in fact an automated script at this very moment is crawling its way across the hard drive automatically backing up all user accounts. The bad news is that, while all the user data is currently intact, it looks like we might have to reinstall the operating system which could result in data loss, but that's why we're backing up the data first before proceeding with any further recovery efforts on the operating system in case -- as all the recovery tools warn -- data could be permanently lost. The good news is that the backup process is over 54% completed, and the custom written script by the Heliohost admins appears to be working perfectly as it cruises along as fast as the hard drive will let it, packages each user up one by one, and ships them off to safety. The bad news is that it will probably take at least another 24 hours to finish the backup. We're talking about literally hundreds of gigabytes of data here, and even though this is Johnny's only current task the bottle neck is of course the hard drive input/output speed and the data transfer rate of the network to send the accounts off. The good news is that we expect Johnny to back online and fully functional within the next week, and it's a possibility that this corrupt operating system problem has really been the root of all of Johnny's increasing number of mostly inexplicable issues over the past couple months. Since we have most of the data backed up already, users whose accounts are located on Johnny can request a copy of their files if they wish by starting a new topic in the Customer Service forum. A link to your files will be PMed to you so random forum readers won't have access to your potentially sensitive information. Once you have downloaded your backup file please let us know so we can delete it. If you also wish to move your account to Stevie you will also need to mention that in your post too because the automatic deletion script doesn't work when a server is offline, but a member of the Heliohost staff can still manually delete a Johnny account even if the server is offline. Our deepest apologies for the massive amount of downtime lately on Johnny, but rest assured that we're all working as quickly as possible to get everything put right as soon as possible.

When it rains it pours, I suppose. Johnny, who always wants to be the center of attention, seems to be at it again. This time cPanel access stopped working for all users whose accounts are located on Johnny, and php seemed to work on some accounts but not others. Also mail service, and FTP also seems to be down for all accounts. At this time we don't have a lot of information to report, but know that it should go without saying that the Heliohost administrators are devoting all of their time towards getting Johnny up and fully functional again. We already have plans that are currently in action that we are hopefully will resolve all of the issues, but it will probably require considerable downtime before everything is fixed. We're hopeful that Johnny will stop throwing tantrums and begin behaving nicely soon.

The problem that caused the massive Johnny downtime over the last day has been discovered, and everything on all Heliohost servers is now online and fully functional. The cause of the downtime wasn't actually one particular thing, but rather a perfect storm of many different small issues that if they had happened one at a time or in a slightly different order would have been easily manageable. However, because of the perfect order in which they occurred, and the fact that they all randomly managed to happen at the same time caused a critical system file on Johnny to become corrupt thus inducing Charlie to go haywire taking down Cody in the process. We came to these conclusions after a thorough log file investigation, and a painstakingly cautious testing progression where each service was brought online one at a time and then meticulously analyzed before proceeding to the next service. The above mentioned critical system file has been successfully rebuilt, and we are confidant that no user data has been lost in the process. Plans are already in motion to prevent anything like this from happening again in the future, but it was probably a one in a million chance for everything to line up the way they did to begin with. After the changes are fully implemented we are sure that the one in a million chance will become a none in a million chance. Thank you for your patience during this process that will undoubtedly leave our servers more stable and more secure than ever.

As you may have noticed Johnny did not immediately come back online after the filesystem scan was completed. In fact during his automated bootup process the exact same unresponsiveness happened again. Recognizing what was happening the Heliohost administrators quickly issued another Charlie reboot to keep Cody online as much as possible while the issue on Johnny is diagnosed and corrected. At this time basic Johnny services are online including mail service, but it will still probably be a few hours before Johnny is fully functional. Please be patient while the admins scour the logs files to figure out what went wrong in the first place. Rest assured that at this time there seems to be no data loss for any of the user accounts. Once again our deepest apologies for the extended downtime. Once Johnny is fully functional another news post will be created to let you all know.

At about 12:10 PDT Charlie crashed causing Cody and Johnny to become unresponsive in the process. With no remote access whatsoever the Heliohost admins quickly contacted a technician in the datacenter that was able to hook up a keyboard and screen to get Charlie working again. Cody, which hosts heliohost.org and helionet.org, was down for about 2.5 hours because of this, and Johnny should also be back up shortly after he gets done scanning his filesystems for errors. The cause of this downtime is unknown at this time, but now that all of the Heliohost servers appear to be functional again preventing something like this from happening again will be our next priority. All user accounts located on Stevie were completely unaffected by this downtime. Apologies to our users for any inconvenience that this may have caused you.

How very astute of you to notice before the news was even posted. See http://www.helionet.org/index/topic/13508-java-upgrade/

Tomcat has been upgraded from 5.5.35 to 7.0.30. Why were we running such a ridiculously old version in the first place? It was because that was the latest version that cPanel offers and supports. We're seriously talking about the cutting edge here, folks. Version 5.5.35 was released July 7th, 2010. That's 2 years, 2 months, and 20 days ago, or 813 days. Version 7.0.30, which we are now running on the Johnny server, was released on September 6th , 2012. That's 23 days ago. It didn't look like cPanel was ever going to update their supported version, so we took matters into our own hands. While we were working on this upgrade it also became painfully obvious that there were way too many accounts with java enabled on them. For instance cPanel warns us in bright red text: Suffice to say we were well over that limit, in fact there are currently 209 topics in the java request forum, and that doesn't include java requests from before we set up that forum or all of the private requests. Obviously, it was causing a lot of issues: High memory usage, high load, tomcat crashes, and various other issues on Johnny were all directly or indirectly related to having too many accounts with java enabled. From here on out please only request java on your account if you really need it, not just because you see we offer it and want your account to have everything we offer. Furthermore, if you find that you were planning on experimenting with java, requested it enabled, never got around to it, and aren't actually using it please make a post on the forum indicating that java service can be removed from your account. Also we will be automatically removing java support from all accounts that become suspended due to inactivity. As a reminder, if you don't log into your account at least once every 30 days it will be flagged as inactive. If you've found that java has been removed from your account while you still need it feel free to request it again. http://www.helionet....-java-requests/ Thank you for your understanding in this matter, and enjoy your shiny new tomcat.

We have a whole mess of software upgrades to announce on the Johnny server: Apache 2.2.23 PHP 5.3.17 MySQL 5.5.22 cPanel 11.32.2 mod_wsgi 3.4 php intl 1.1.0 php fileinfo 1.0.5 suPHP Some of these services were still running really old outdated versions and hadn't been updated in literally years. MySQL, for instance, was still running version 5.0.96. The Johnny server seems to be running a lot more smoothly since the upgrades, but only time will tell if this will have a lasting effect on the uptime and stability. A lot of the upgrades that were performed were actually requested by members of the community so be sure to make a suggestion if you think there is something missing. php intl was requested here, php fileinfo was requested here, and MySQL 5.5+ was requested here. Thank you to all the members of the community who have taken their time to make a post requesting an upgrade. It really helps to have your feedback on prioritizing what needs to be done and in what order. As always Johnny is our more cutting edge experimental server so of course it's going to get upgraded to the newer versions first. Stevie, as is quite evident through the uptime images below, is our stable server focused on maximizing uptime. Not only do new versions need to be tested thoroughly worldwide, but they also need to be tested thoroughly here at Heliohost to see how they perform on our systems. Once we are satisfied that these upgrades won't adversely affect the stability or uptime of Stevie, he will receive a similar treatment.

Please join me in congratulating the newest member of the Heliohost support staff, Ice IT Support. We get the question quite often, "How do I get promoted?" and the answer is actually quite simple: Just help people. That's exactly what Ice IT Support has done. At Heliohost we have a different business model than most companies. We believe in building and strengthening our community by encouraging users to assist each other with questions and problems. Not only does it free up the admins to spend their time working on bigger issues, but it also promotes a great atmosphere of learning and helping each other. You may think you already know a lot about servers, websites, and computing in general, but it's amazing how much more you can learn trying to help others with their seemingly easy problems. After proving himself by helping numerous people on the forums Ice IT Support drew the attention of the existing moderators and administrators of Heliohost. After some internal discussion we offered him the position and he accepted. Congratulations Ice IT Support, it's because of users like you that makes Heliohost great, and ensures that this free service will exist for many years to come.

This is a great time for all Heliohost users to think about their own account and website security since there have been a few websites hosted on Heliohost that have been hacked recently. First of all server security has been analyzed and there does not appear to be any vulnerabilities and the servers have not been compromised in any way. As always Heliohost takes the security and privacy of our users accounts at maximum priority, but we also believe in giving our user the maximum amount of access. Most free webhosts restrict the capabilities of accounts so much that they are barely usable, but here at Heliohost we try to do the opposite. We try to give users as much access as possible. Since we don't restrict everything down to unusable levels it just means you have to keep your own website security into consideration. Don't worry, most of the steps that make the biggest difference as really simple common sense things, but never forget that each user is responsible for their own account and everything located on their account. Choose a strong password. I recommend 10+ characters with uppercase, lowercase, numbers, and symbols. You can use this website to generate strong passwords. Don't use admin username. If you use a CMS of some sort such as Wordpress, Joomla, etc don't use the default username to log into the administration area, and as always choose a strong password. Use secure connections. I know a lot of people probably love to sit around at coffee shops blogging about how inept the barista is, but public wifi can be really insecure. Also if you want to purchase and install SSL on your website it will make it even more secure. Learn unix security. Unix OS can be very daunting for new users, but getting all your permissions correct can be a huge boost to your site's security. Learn Apache security. Using .htaccess files, password protecting directories, placing index files in every directory, disabling directory indexing, etc can all increase your security too. Use latest CMS versions. Keeping your software up to date will go a long way towards preventing your site from being compromised. The longer you put it off the more vulnerabilities in the version will be discovered and the more script kiddies will start exploiting it. Disable CMS version reporting. Similar to the above point if your site is announcing its version it just makes it that much easier for hackers to know what exploits it's vulnerable to. Only download from trusted sources. A lot of free themes and plug-ins are actually infected with backdoors allowing hackers access to your site. Once you install the theme/plug-in and your site is publicly accessible it will check in with its creator and it's only a matter of time until that hacker does something malicious to your website. Take backups often. If your site does end up getting messed up somehow being able to restore a backup and only losing a couple days (or a couple hours) of content will save you a ton of grief later on. Another thing to keep in mind is that every hacked website we have investigated recently has been a wordpress installation. Wordpress is a really popular CMS, and isn't necessarily insecure, but because of its popularity it's vulnerabilities are widely known. We definitely don't intend to discourage anyone from using Wordpress, but you might want to read through this article http://codex.wordpress.org/Hardening_WordPress if you use wordpress and you're concerned about your site's security.

The long-postponed, oft-dreaded, major software upgrade on Cody has now completed, and load levels should be returning to normal. Unfortunately, this software upgrade has had a greater effect on the other servers than we expected. Johnny had high load and intermittent downtime during the entire process, and even stable-as-a-rock Stevie experienced higher than usual load and a little slow down. We apologize for the small downtimes on Johnny user accounts, but this is why news was posted in advance so that all Heliohost users would hopefully be ready for the unexpected. With the software upgrade finished and in place your accounts and our systems should be more secure than ever. Thank you for your patience during this process.

Starting on July 27th at about 15:00 UTC or 08:00 PDT, the Cody server will be undergoing some major software upgrades in order to increase stability of the entire Heliohost server cluster. Cody is the server that houses our homepage heliohost.org, our forums helionet.org, and our wiki wiki.helionet.org as well as one of our two nameservers, and the central account database. While every attempt will be made to keep all services online during the upgrade, there will undoubtedly be higher load than usual and some service interruption may be inevitable. No user accounts will experience any downtime from this upgrade because user accounts aren't located on Cody, but new account creation may be temporarily suspended. If you normally login to your Heliohost account from our homepage it may be useful to know that you can also login at http://stevie.heliohost.org:2082/ or http://johnny.heliohost.org:2082/ depending on which server your account is located on. Also, if the support forums are slow or down due to the maintenance keep in mind that support is always available on our live chat channel. If there are any volunteer admin available to help you they will be glad to. Thanks for your patience during this major software upgrade.

On a more personal note, this should probably be common courtesy to most of you out there, but camping out on our free service trying to leech people onto your paid hosting provider during one of our rather short and infrequent downtimes is not just against our terms of service, it's just ridiculously sleazy. We understand that Heliohost is an amazing service for free. That's why we're here in the first place helping out and making it better. Before this downtime Stevie had a weekly uptime of 99.80% which is better than a lot of paid hosting services. Sorry to be stating the obvious since this is a complete no-brainer for most of you decent people out there, but apparently certain folks don't realize something is wrong unless you explicitly point it out to them.

At approximately 08:34 UTC an automatic process on Stevie malfunctioned causing corruption in one of the main server configuration files. This caused all user accounts located on the Stevie server to display account queued pages or 404 file not found errors. As soon the corruption was discovered the Heliohost administrators immediately began work to fix the problem. At approximately 12:04 UTC this process was completed and service was restored to normal functionality. Accounts located on the Johnny server were completely unaffected by this downtime, and heliohost.org and helionet.org continued to operate normally. Thank you to all the Heliohost users who reported this problem and assisted in diagnosing the issue. With the assistance of all the reports we have received we now know what caused this to happen in the first place, and rest assured that steps are being taken to prevent it from happening again in the future. You may also be interested to know why this announcement wasn't published any earlier. Here at Heliohost we believe in taking action first and talking about it later. Rather than responding to forum, chat, and twitter questions our time is better spent first fixing the problem, verifying that everything is working again, analyzing why it went wrong in the first place, and taking steps to prevent it from happening again. Then after all that is said and done then we spend some time making announcements and talking about it. Thank you for your patience during this process. If you are still seeing a queued page on your account make sure you clear your browser cache, and if that doesn't fix it feel free to let us know by starting a new post. If you're having any other issues please don't hesitate to let us know about those too.

All Johnny services are back online, and the high load issue should be returning to normal. Thank you to all of the Heliohost users who have reported the various issues they have been experiencing. The more information we have about what is happening the better we can analyze the situation and take steps to prevent these types of incidents from happening in the future. Once the load drops down to a reasonable level we should be able to restart new account creation and deletions soon. Thank you for your patience during this process to make Johnny more stable.

Probably related to http://www.helionet.org/index/topic/12830-johnny-high-load/

Heliohost's Johnny server is not unfamiliar with bouts of high load. In fact, since that server offers the full Heliohost spectrum of services it is expected to be a little less stable than Stevie, but for the last twelve hours or so Johnny has been even less stable than usual. The Heliohost admins are looking into the cause of the high load, and working on ways to correct it. The servers generally operate best if the load is less than 20, while less than 5 is optimal; at the time of this post Johnny's load is averaging around 150. Usually the best course of action when load is too high is to disable account creation, deletion, inactivity suspensions, and other non-essential high load processes to simply wait for the load to return to a reasonable level on its own, but today's case of high load is a little more severe than that. Please pardon our downtime as we work to get Johhny's spiraling out of control, high load back into a reasonable level.

At approximately 13:38 UTC there was a disruption in the internet connection between 10gigabitethernet1-1.core1.fmt1.he.net [72.52.92.109] and lafrance-internet-services.gigabitethernet3-15.core1.fmt1.he.net [66.220.10.126]. The cause of the disruption is still not clear, but it blocked all requests to any of the Heliohost services including our home page, the support forum, and all user accounts. Since this disruption was not caused by any of the Heliohost hardware we were not in direct control of the faulty connection and there was nothing other than contacting the responsible parties that we could do. Fortunately service was returned to normal swiftly, and we hope that this will not occur again. Internet service was restored to our servers at approximately 14:28 UTC.

The webmail client Roundcube has been re-enabled on Stevie server. It was originally disabled over a year ago during one of Stevie's bouts of bad performance. Roundcube webmail process was found at the time to be using a considerable amount of memory and cpu time so it was disabled in an attempt to improve the stability of the Stevie server. While it may have helped Stevie out some over the last year, the Heliohost administrators have decided to give it another try. We of course always aim to keep Stevie as stable as possible, maintain as high of an uptime as it is possible to, and provide the very best quality of services to you. Roundcube has always been available on our Johnny server, but if you have an account on Stevie you can help us test the functionality by logging in to one of your Stevie hosted email account via this link http://stevie.heliohost.org:2095/3rdparty/roundcube/. If you discover any issues with the newly re-enabled Roundcube, please don't hesitate to inform our admin team by posting a new topic in our Customer Service forum.

It seems a larger number of people than usual are having issues logging in to their Heliohost accounts recently. Please take note and remember a couple of commonly made mistakes when it comes to logging into your account: Usernames are case sensitive. If your username is 'joeshmoe' and you type 'JoeShmoe' it won't work because of the capital J and capital S. Passwords are case sensitive. Likewise, if your password is 'pAsSwOrDz123' and you type 'passwordz123' it also won't work even though you technically have all the characters correct they are not in the correct capitalization. Max username length is eight characters. Even if you would rather your username be 'joeshmoevindiddlelinski' you'll have to find a way to shorten that to eight characters or less, such as, 'joeshmoe' which is of an acceptable length. Usernames are all lowercase plus numbers. Even if you register with the account name 'JoeShmoe' the system will convert this to lowercase for you and the registration email will say your username is 'joeshmoe'. Also be careful to make sure you're using the correct username and password combination before you hit submit, because we have brute force protection enabled on our servers. If the system detects you guessing the wrong login information multiple times you will be disallowed from logging in for a period of time. If you're still having issues logging into your account feel free to create a new post asking for assistance.

See http://www.helionet.org/index/topic/12743-server-partition-usage/