SAS Enterprise Guide 8.2 Update 2 - Server Library List Problem

[SASKiwi]

I just installed the latest update today via the EG Check for Updates. When expanding the Server Library List, the wheel spun for a while then EG locked up. The only way to close it was to kill it with Windows Task Manager. I tried this several times with the same result. I reinstalled the original 8.2 to get around the problem.

Just wondering if anyone else has experienced similar issues.

[JuanS_OCS]

Hey @SASKiwi ,

thanks a lot for the flag. I just did a local update of my EG, mine works OK, so far.

Let us wait for other experiences, as I wonder as well.

[Technolero]

I installed the update on three different computers and have not experienced the issues you described.

[SASKiwi]

@Technolero , @JuanS_OCS - thanks for the feedback. I'll probably raise a track on this as I have an EG application log that might point to what is going on.

[Technolero]

Are you connecting to your servers via VPN? Since COVID the majority of staff at my organization are working remotely via VPN, and during peak usage VPN speeds drop so low that EG responsiveness really suffers, almost making it unusable. I now make sure my EG sessions are complete before most of the users log in for the day.

[SASKiwi]

@Technolero - Not in the case of this issue. We did suffer that problem some time ago when working from home until VPN was optimised.

[_echo_off]

I have noticed similar behaviour. But I haven't figured out what it is yet (we have assigned a lot of Hadoop libraries in that case)...

[SASKiwi]

@_echo_off  - Appreciate the feedback. We too assign a lot of libraries, although not Hadoop, so I was wondering if it is the number of libraries being assigned that could cause this. I'm hoping to have more time to check this out in the next few days.

[_echo_off]

I've run some more tests now. In my case the number of libraries is more than 650, which works without problems in EG 7.15 HF7 and 8.2 HF1 (result within seconds). EG 8.2 HF2 freezes and disconnects from the SAS server after minutes of inactivity. I have now generated log files and will open a corresponding track at SAS...

[_echo_off]

The problem may be in the "AttachToLibraryState" function. The log file for HF2 is full of related messages (unlike HF1)...

[_echo_off]

The problem has already been reported by other customers too and has been forwarded to SAS Global Support...

[SASKiwi]

@_echo_off - I have also opened a track and got the same response that it is a known problem. We have about 130 libraries. I'm told that EG eventually unfreezes but you might have to wait for up to 30 minutes.

[ccaero]

Hi @ all,

i have noticed the same - we have about 120 preassigned librarys and expand the librarys in the server-browser takes around 15 minutes via VPN (homeoffice) and around 5 mins at our main-office.
As a workaround you can open/browse librarys via the" file -> open"-menu of enterprise guide.

I think EG 8.x series has generally an performance-issue at the server/library-browser. With EG 8.2 update 1 the library expanting nearly so fast as EG 7.15, but opening tables takes around double time compared against EG 7.15.
In EG 8.2 update 2 the tables opens very fast, but instead the library-browser expands very slow compared against update 1. Maybe they had moved some loading-work from table-opening-event to library-loading.

I think this is related to displaying the variables the tables in the library-browser. The support tells me, that performance will be better with EG 8.3 (release date around mid august).

[JuanS_OCS]

Good insights, thank you @ccaero 

[SimonDawson]

I'm the engineer helping @SASKiwi with his support track. We have successfully replicated the problem here at SAS and are investigating.

In my testing I had to wait up to 3 minutes for my lab system, not 30 minutes. Sorry about the typo in my email @SASKiwi .

The length of time in the pause is proportional to a few factors from my experimenting. Network latency and number of libraries are the main factors by the looks of things. If you have long ping times to the workspace server and many libraries the time the UI will block will be longer than if you have short network latency to the server and fewer libraries.

I'll do my best to keep the community updated with what we find in our investigation.