Should the web interface be this slow?

Hi,

I'm having trouble with the DSS web interface. It takes between 20-60 seconds to load each new page. Sometimes, the interface freezes altogether and it takes up to 10 minutes before the machine becomes accessible again.

To make things even worse, each time i bring up the event window by clicking at the flashing exclamation mark (i get some cmnd_abort(1143) errors from an iSCSI target) and click on one of the errors to get some more information, the whole interface locks up totally. The page in the window that opens is never loaded and the rest of the interface stops working.

So, the big question: what did I do wrong?

Any help is appreciated!

Kind regards,
Soren Petersen

(I'm currently using the 16TB DSS version. I had the same problems with the free 1TB version of open-e)

@Soren: I don't think you do anything wrong. It is the same behavior at my site with 24TB total size. (I thought it would be my problem only, as I could not find any other thread about it ...)

Poka,
Korlojow

G'day Soren,
Sometimes this happens to us as well, I had put it down to problems with the Adaptec Raid controller, as we had no issues with the DSS Lite running on a PC Motherboard and software raid.
It seems to be mainly worse when the controller is rebuilding.... But that is only my untested gut feeling. Since the other services (iSCSI etc) work fine, we have just lived with it.
I too thought it was just our problem, didn't realise it was common.....

Rgds Ben

Quote:

---

Originally Posted by sorenp

Hi,

I'm having trouble with the DSS web interface. It takes between 20-60 seconds to load each new page. Sometimes, the interface freezes altogether and it takes up to 10 minutes before the machine becomes accessible again.

To make things even worse, each time i bring up the event window by clicking at the flashing exclamation mark (i get some cmnd_abort(1143) errors from an iSCSI target) and click on one of the errors to get some more information, the whole interface locks up totally. The page in the window that opens is never loaded and the rest of the interface stops working.

So, the big question: what did I do wrong?

Any help is appreciated!

Kind regards,
Soren Petersen

(I'm currently using the 16TB DSS version. I had the same problems with the free 1TB version of open-e)

---

We're using DSS on a dual-CPU server with a total of 8 cores, have none to light load on the DSS (CPU usage below 10%, load at 2.0), I/O traffic on the disks and the USB module at a "why did you buy so expensive hardware and don't us it?" level, the LAN is free as a highway at night, but it takes up to 30 seconds to switch between FC group displays even when not having done any change that has to be applied... when there is a change, then it gets even worse.

It's like having the parking brake on and no license key to release it. There must be something wrong - with the system design.

Good to see we're all in the same boat - have a look the this message (http://forum.open-e.com/showpost.php...7&postcount=2), somebody is working on the ui, so voice up!

With regards,
Jens

G'day Soren,
We updated to the current version of DSS(2nd Nov?) yesterday and so far the response time has been vastly improved for us. (Even with the degraded array).
Are you running this version?

Now our only problem is we are back to slow Network speeds. (<10Mb/s) but that is another topic...


Cheers Ben.

Hi,

This is a first but I have the same issue here.

I'm running:

Model: Data Storage Server
Version: 5.0.DB49000000.3278
Release date: 2008-11-19

Normal I/O very low CPU usage, almost idle network as I'm accessing from a management interface which is idle most of the time (no NAS or iSCSI traffic here).

NAS and iSCSI speeds are OK (the same 10MB/s "barrier") its just the GUI that is slow as hell. (btw, it took me about 4 min to get version to paste it above).

This was running at normal speeds about 4/5 weeks ago. But lately it is impossible to use. Uptime is about 110 days no. Never been this long before.

I'm not going to reboot it just to see if the GUI gets faster, I have a lot of systems using iSCSI targets that will switch to read-only if there is an iSCSI disruption and that will be a big mess fix latter.

Any ideias?

Sounds like something in the backend could be causing this... can you send the log file to support@open-e.com so we can see what might be happening. Are there any event error from the GUI?

Hi HangaS

Check the raid to see if it is ok (no bad disk)
also check to see if there is a raid verify going running

don't forget to send the logs to support

Hi Todd and symm,

From what I can see in the 3ware 3DM2 web interface (which responds fast) there is nothing wrong with the raid array, and no rebuild was done since February.

There are no events in the GUI also that I can see, but I don't know if they don't show because of an HTTP time out. The event window loads after a few minutes, but 0 events.

I will send the logs to support as soon as I can get them. I'm trying to reload the Status -> hardware page for the last couple of hours.. without success.

I guess the GUI might be an apache or similar running some application, is there anyway of restarting the web server without having to restart the whole system?

Thanks!

I contacted support with the log files and I was told that there were detected problems with the filesystem.

Now I ask, which file system might this be? the FS where the NSS resides on the flash disk or the filesystem of raid array itself?

I was told to run the Repair Filesystem under the Extended console tools, but this looks to me that it will repair the LV where all the data is on.

How could a problem in this filesystem affect only the GUI and not the volume performance, as everything appears to be OK with the volumes?


Anyway, it I'm to do this, how long does it take to repair a 2TB volume? Cos I have to schedule a maintenance window for this, as I will have to reboot the DSS.


Kind Regards,
Miguel

The Filesystem repair will be for your Logical volumes not on the module as the module runs from memory. Engineers have asked you to run this and make sure to verify a memory test. To do this reboot the system and after POST enter the ESC or TAB key to access a basic menu.

For the time on the repair for 2TB depends on many factors but anywhere from 1 - 7 min. if all is running smoothly longer if there are other issues.

so is this slow UI issue resolved?

I am going to try out open-e with Adaptec RAID & Intel SSDs. What is the best filesystem to use in order to prevent all these "delay"?

Yes, the issue is solved. Apparently it was related with remote LDAP auth.
Although the network was correctly configured and that the LDAP server is working well with other systems, somehow disabling the LDAP auth brought the system back to normal speed.

I decided to disable LDAP for now, as I was not using it lately.