Yesterday, August 24, 2010 at approximately 3:48PM we experienced a brief network outage for approximately 16 minutes. This issue is a result of a JunOS bug causing recent problems with one of our Juniper distribution stacks. As such, we are scheduling a maintenance window for this weekend, Sunday August 29th 01:00-04:00 EDT to make the necessary software upgrades. We expect the maintenance window to be no longer than 30 minutes, however if there are complications, we are reserving this 3 hour window for the maintenance.

While the outage caused today was related to this issue, we do not anticipate another outage to occur before the scheduled maintenance window. We appreciate your patience and thank you for your understanding that the matter is and will be resolved this upcoming weekend.

The VPS1 node was replaced earlier this month but the new server didn’t come built with the requested memory (18gb) and only has 12gb. The node is going to run ALOT faster with the extra memory and we may further increase this to 24gb if needed. For now we are going to increase to 18gb Sunday 22nd August at 3am so please allow an hour for this maintenance window. The actual upgrade only takes 20 minutes but it takes a while for all the containers to come online.

UPDATE: Unfortunately we were unable to proceed with the upgrade as we didn’t have the correct memory in stock. This will be rescheduled for Friday night.

Recently Atom server was replaced with a new machine and more memory (4gb total). However the new 64bit architecture allows CF to use more memory which is great for CF performance but not so great if the server doesn’t have enough memory. So on Wednesday early morning at 3am we will increase the server’s memory to 6gb total. Please allow 30 minutes downtime for this maintenance window.

UPDATE: Memory upgraded, no issues, enjoy the added performance!

Please be aware that only persistent server issues are posted here and normally 30 minutes after an incident occurs as techs first need to check the issue before posting an update.

At 3am August 3rd (EST) we are going to replace the current VPS1 node with a new server and more memory. All we need to do is take the RAID 10 array with its 4 drives out of the old server and place everything into the new server. It takes a while for all the containers to start so please allow an hour for this maintenence window.

We have some kind of network issue with a few racks of servers down. We are looking into this.

UPDATE: We are still waiting to hear from teh data center as to what the issue is.

UPDATE: Looks like most servers are coming back up now.

UPDATE: Three of our switches just stopped responding. We have power cycled them and the servers are responding now.

Apologies for this mysteriously outage. We are going to investigate the cause.

We will be upgrading Plesk on Neutron to 9.5 on Tuesday July 20th 12am which is basically Monday night.

This is a significant change as Plesk 9.5 looks somewhat different to Plesk 8. Version 9 has been out for over a year and we have it on all new Linux servers and it’s now time to deploy across all Windows servers.

Plesk 9 allows Overselling so no longer will clients have to juggle their disk space and bandwidth allocations and it also comes with the new Reseller level where you can create Client Accounts and assign multiple domains to each. Until now resellers have been limited to assigning client access per domain only. I believe the default skin will be a Vista type skin but you can change this to a more familiar XP skin (or whatever type skin you’ve been using).

It’s a big upgrade so there might be a few issues afterwards. I urge everyone to check everything after the upgrade to make sure things are in the right place. As always please do submit support requests if you find issues.

We estimate about 2-3 hours for this upgrade. The upgrade was completed on Baryon today and it went extremely smoothly.

Any resellers that want the new Reseller level please email sales@hostnexus.com after the upgrade and I’ll make the changes.

UPDATE 8am: The upgrade is almost done and the server is up and serving up websites. The upgrade process is running through all the mailboxes and checking permissions and has been on this stage for 2-3 hours. We just have to let it run its course and mail is stopped while this happens. On our other upgrades this step normally completed in under an hour so we are not yet sure if this indicates an issue. After this stage is complete the internal Plesk database is converted to 9.5. We have no ETA as we cannot see any progress bar. But if it continues on much longer we will bring in the Parallels engineers to take a look.

UPDATE 22nd: The process did complete by itself and mail came back up. However, the upgrade wasn’t a full success and the Plesk control panel is a bit of a mess. We have the Plesk engineers working on that.

Everything else green across the board.

We will be upgrading Plesk on Baryon to 9.5 on Monday July 17th 12am which is basically Sunday night.

This is a significant change as Plesk 9.5 looks somewhat different to Plesk 8. Version 9 has been out for over a year and we have it on all new Linux servers and it’s now time to deploy across all Windows servers.

Plesk 9 allows Overselling so no longer will clients have to juggle their disk space and bandwidth allocations and it also comes with the new Reseller level where you can create Client Accounts and assign multiple domains to each. Until now resellers have been limited to assigning client access per domain only. I believe the default skin will be a Vista type skin but you can change this to a more familiar XP skin (or whatever type skin you’ve been using).

It’s a big upgrade so there might be a few issues afterwards. I urge everyone to check everything after the upgrade to make sure things are in the right place. As always please do submit support requests if you find issues.

We estimate about 3 hours for this upgrade. This is the first Plesk 9 upgrade on a live Windows server so we are allowing more time than normal. The upgrade has been tested on our test server and things went very well so I’m not forseeing any major issues.

Any resellers that want the new Reseller level please email sales@hostnexus.com after the upgrade and I’ll make the changes.

UPDATE: The upgrade went extremely well. The actual upgrade was done in under an hour and then we did some clean-up of some errors for another few hours. But overall, very smooth and minimal interruption. We hope they all go this well!

As always please do check everything and let us know right away if you have any issues so we can check and fix. Also don’t forget to contact us if you want the new reseller level so you can create Client Accounts.

We will be upgrading MailEnable on Boson to MailEnable Pro with its IMAP and TLS/SSL support on Saturday 17th July 12am US EST. This is normally a 3-4 hour upgrade during which time mail is queued for delivery and outbound mail is temporarily stopped.

Many thanks for your co-operation.

Everything else green across the board.

UPDATE: This went well and took about 5 hours. We just need to reboot the server which we are doing now.