Skip to main content

vmProtect time zone issue

Thread needs solution
Regular Poster
Posts: 71
Comments: 162

After having so many problems with vmProtect 8, along with suggestion from Acronis, I deployed another VMP 8 virtual appliance in my vSphere environment. I renamed the original AcronisAppliance1 and deployed the new one as AcronisAppliance2.

I recreated half my jobs on the new appliance and then deleted those jobs on the original. This is in hopes to distribute the load and get more consistent backup successes.

I noticed when adjusting times in the original, old appliance, the display time on the task screen where it says Next run time was different than what it said if you go into the job and edit it. So I did check the console of each one and I've attached the screenshots. We are in GMT -5 Eastern Time (US & Canada). So the new appliance just deployed today on the right hand side of my screenshot is correct. The only way to get the console to display the correct time on the old applince (on the left) is to select GMT -01:00) Azores. I don't even know what Azores is, but I will wikipedia that after this post! If I choose GMT -5 Eastern Time it makes the time 11:01 AM (as of this writing I just tried it).

In my second screen shot you will see that the task screen shows a job to run at 12:30. However if you go into EDIT on that very same task, it shows it is to run at 17:30 (which is when I want it to run). So what time will the job really run?

Why is that?

0 Users found this helpful
Acronis Program Manager
Posts: 22
Comments: 3670

Hi KJSTech,

The "Next run" (12:30) value is calculated basing on the time-zone settings of the machine where you run Web Console (Windows) and the time settings of the Virtual Appliance. The 17:30 time is the virtual appliance time, so when it is 17:30 on _virtual appliance_ then the task will run. The virtual appliance originally takes time from the ESXi host (Configuration->Time Configuration in vSphere client) and then adds the time zone shift to this time. The time on ESXi host is stored in GMT without offset and timezone offset is added by vSphere client itself taking it from system where vSphere client is run.

It looks like that the 1st appliance has been deployed from the machine which had GMT-1 time zone (this time zone is automatically injected into virtual appliance during deploymet), while the 2nd appliance was deployed from machine which had GMT-5 timezone.

In other words you should check the timezone settings on the machine (Windows) where you run the web console from as the time values you see in UI highly depend on this.

See also screen shot in the attachment: I've changed the time zone on appliance to GMT+3, while on my machine it is GMT+4 and current time is 10:36. The task is set to start at 17:30 (GMT+3 appliance time), so when on appliance it is 17:30, on my machine it will be 18:30 and the task will start at 18:30 GMT+4.

P.S. What concerns your original issue with stuck tasks we've found that the problem is most likely caused by the network share specifics (the backups location) and how it is handled by the appliance Linux kernel - there are CIFS errors (in /var/log/messages) which are retried but do not fit into the Linux OS timeouts which therefore causes the hanging, so indeed splitting the jobs should do the trick.

Thank you.
--
Best regards,
Vasily
Acronis vmProtect Program Manager

Attachment Size
129622-107398.png 143.25 KB
Regular Poster
Posts: 71
Comments: 162

Ok my machine is where I deployed both appliances to vcenter. vcenter and all the ESX servers, as well as all windows servers, network equipment, phone system, everything... sync using NTP using 2 domain controllers (primary is a physical stand alone server, secondary is a virtual).

I checked the time zone on my machine and it is GMT -5 (EST).

I don't know why I can't change the time zone on the virtual appliance 1 to GMT -5 (EST) and have it look correct.

Anyway the backups last night went well after splitting them up. I only had 1 failure about snapshot quiesced timeout on the new deployed AcronisAppliance2. I was unable to even snapshot that machine in VMWare. Did some reading and changed disk.enableuuid to false in the vmx config for the machine and now the backup for that one particular case runs successfully.

The old appliance writes to one CIFS server (EMC Celerra NX4) and the new appliance writes to a Buffalo Terastation NAS unit. So the backups are not all writing to the same place now either and hopefully that lessens the load on the EMC Celerra NX4.

I also have an old server running Openfiler NAS and I haven't tried backing up to it yet. I always have to make sure I have available LOCAL accounts to our NAS / CIFS devices because domain authentication does not work due to NTLM level 5 (Send NTLMv2 response only\refuse LM & NTLM) which I think is a SAMBA issue if I recall correctly.

Regular Poster
Posts: 71
Comments: 162

If I break out to the command line on our AcronisAppliance1, and run the date comand, the time is correct 09:17 as of this screen capture. However if I switch back to the gui, its wrong!

I changed the timezone in the GUI to GMT -5 which is should be and then powered off the VM and powered it back on... thinking it would re-read the correct time during boot.

Is there any way to set an NTP server?

Attachment Size
129638-107404.jpg 31.48 KB
129638-107407.jpg 82.16 KB
Acronis Program Manager
Posts: 22
Comments: 3670

Hi KJSTEch,

The appliance is synchronizing the time via embedded VMware Tools with the ESXi host itself (where appliance is deployed to) and via 'date' command its showing UTC (GMT) time without timezone offset which is added in UI (-5 hours in your case). Therefore it is using the NTP server used by the ESXi host itself. Can you check the time settings on the ESXi host? What does it show there? See screen shot in the attach.

Thank you.
--
Best regards,
Vasily
Acronis vmProtect Program Manager

Attachment Size
129641-107410.png 101.07 KB
Regular Poster
Posts: 71
Comments: 162

The ESX server (all three of them actually) are configured the same way. They look at two NTP servers. The local time appears correct on each ESX server. So no matter which ESX server AcronisAppliance1 is vmotioned to, it should be correct. Note: newly created AcronisAppliance2 does not exhibit this problem.

Screenshot of our NTP config off of the server AcronisAppliance1 is currently running on.

Attachment Size
129649-107413.jpg 109.91 KB
129649-107416.jpg 47.77 KB
Acronis Program Manager
Posts: 22
Comments: 3670

Hi KJSTech,

I've just double-checked with our developers the timing issue: the base UTC (GMT) time is taken by virtual appliance from the ESXi host, however this is not done via VMware tools by default (ESXi can throw the time through BIOS to the guest VM). In order to enable the time synchronization you should enable corresponding checkbox in the appliance settings in vSphere client (under Options->VMware Tools). See the attached screen shot. This should synchronize the time on the appliance with the host forcibly.

Thank you.
--
Best regards,
Vasily
Acronis vmProtect Program Manager

Attachment Size
129840-107449.png 43.42 KB
Regular Poster
Posts: 71
Comments: 162

OK I selected that option but it still does not make GMT -5 (EST) the correct time. I tried rebooting and also powering off/on the virtual appliance.

I think I might have to scrap that virtual appliance. Remember with all the trouble tickets we decided to spread the load across 2 virtual appliances? Well the newly deployed second appliance has no issues (time zone, or otherwise). The original old appliance that's been with me for a long time now had 2 failed backups this weekend. It just seems to be a problem child across many areas from time zone display to backing up systems.

Since there is no way to export my jobs, I will spin up a new appliance when I get the time at some point this week.

Acronis Program Manager
Posts: 22
Comments: 3670

Hi KJSTech,

Yes, you may be right and the issue is with the particular appliance instance (possibly the result of the previous corruption).

Just as a tip: here's the thread where I explained how to migrate settings from one appliance to another (in vmProtect 9 we'll have this feature in UI, but until then this method should work): http://forum.acronis.com/forum/34725#comment-107881

Thank you.
--
Best regards,
Vasily
Acronis vmProtect Program Manager

Regular Poster
Posts: 71
Comments: 162

Ok thanks for the tip. I deployed a new virtual appliance. I followed your tip and all the jobs followed, including the job history of fails and passes.

Though the console is now correct with the time (GMT -5 - EST), the view tasks screen shows the wrong "next start time". If you edit the task the time is correct but just on the main page its wrong (just like the old appliance). Maybe it would have been a better idea to recreate 18 jobs from scratch along with their customized subject line email on failure. But I will let it go and see tonight if these backups work ok.

The new appliance I created last week is fine. The view tasks screen next run time matches the schedule if you were to edit the task. But I manually entered those 16 jobs by hand one by one.

Regular Poster
Posts: 71
Comments: 162

Vasily,

I migrated all the tasks to the new appliance, however those tasks are not running as scheduled. They didn't seem to run at all (no success OR failure). I was able to kick one off now.

So I went through each job and did an additional "SAVE" on it... maybe that will "reset" it somehow? I'll see if the jobs run as scheduled tonight. They are scheduled fine, but just didn't seem to kick off.

I followed your procedures exactly for moving tasks from one appliance to another using WinSCP.

Regular Poster
Posts: 71
Comments: 162

Confirmed... scheduler broken on new appliance.

A job was supposed to kick off at 17:30. Its now 17:35 and it never kicked off.

Can you help me with this?

EDIT: I altered the job to run at 17:40. When I saved it, the last run time changed to N/A on the tasks list. 17:40 came and the job started.

So I went through all my jobs... (there are 18 on this virtual appliance) and advanced the start time by 5 minutes (no big deal). All of the last run time's changed to N/A (from the previous last run time when they were on the old appliance I imported these jobs from).

So I think the jobs will run now. I'll know tomorrow morning when I come in and check on the appliance.

Acronis Program Manager
Posts: 22
Comments: 3670

Hi KJSTech,

Typically there is no need to update the tasks schedule after such manual migration of tasks, though indeed editing the task and saving the new slightly different schedule can refresh the scheduling if something goes wrong. In any case please let us know how it goes over night, i.e. whether the jobs run properly on new appliance.

Thank you.
--
Best regards,
Vasily
Acronis vmProtect Program Manager

Regular Poster
Posts: 71
Comments: 162

Ok for the record using your procedure to migrate all scheduled tasks from one virtual appliance for another half worked. All tasks job definitions were imported successfully. The problem is that even when the time hit for a task to run, no tasks ever ran.

So after modifying all tasks to start just 5 minutes later, last night they all ran successfully.

Yes every task scheduled last night ran. Even the problematic mail server backed up successfully!

Thank you for your help. Just changing the time for each task after importing them was a lot easier than manually designing each task from start to finish.

Just a heads up for AVP version 9 because I know you said that exporting and importing tasks was considered to be a new feature. I'll make sure to test that if there comes a time for a public beta of 9!

So as far as the original topic goes, the newly deployed virtual appliance has the correct time zone (GMT -5 EST) out of the box, with no additional configuration needed. Task list in the web UI now list the correct time on both the list AND when editing the job's properties.

Who knows what was wrong with the old virtual appliance... but I have it powered off and at some point I will likely delete it.

Forum Member
Posts: 0
Comments: 62

So how did this work out for you? Would be nice to get some updates here.

Beginner
Posts: 0
Comments: 1

I just wanted to say that the screen shot of 'sync guest time with host' will fix this for 99% of the people who visit this thread. Don't know what the issue dude was having exactly was, but that fixed my issue.