Re: elogd hangs, posted by Alan Grant on Fri Aug 18 16:39:21 2017
|
Okay I will just try the debugger approach first and then take it from there. Thanks Stefan.
| Stefan Ritt wrote: |
|
Well, having the config and data files only help if I can reproduce the hanging. If if you give me your files and a step-by-step instruction on how to reproduce the hanging, I can give it a try. But if it happens randomly after a while, it will be very hard for me to reproduce and fix it without the exaclty same user access pattern which of course I don't have.
| Alan Grant wrote: |
|
I could begin debugging with C++. In the interim, if you think it will help, I can also provide you with my Config and log files, and the section of redacted data encompassing the time of the hang - just let me know. Regarding CPU usage, I have noted that in the past and have never seen very high CPU usage during an Elog hang. The VM itself remains responsive.
| Stefan Ritt wrote: |
|
I have to figure out where elog hangs. I guess it must be some kind of endless loop, triggered by some corrupt data in one of the elog entries. Under linux this is fairly simple (just run elogd under the gdb debugger, wait until it hangs, then press ctrl-c and enter "where" to see a full stack dump where elogd is currently executing). Under Windows this is more difficult, since you need Visual C++ from Microsoft to do the debugging. One thing you can do however without VC is to check if the CPU time is consumed to 100% by elogd, indicating an endless loop.
Stefan
| Alan Grant wrote: |
|
I have a very long standing problem with elog over the last few versions where almost daily the service will hang. Cannot even Restart elogd, that just hangs. Clients experience Page not Found. I can only get the service reinitialized by rebooting the VM machine. I have Elog verbose logging On plus a number of external triage monitors running but nothing is yielding clues beyond the precise time the hang occurs. Aside from providing the Config and log files what else can I provide for you to assist, and what other triage measures can you suggest I try? FYI, there can be up to 20 users at one time doing searches (not updates), and I've trimmed the depth of log files that can be searched so that the CPU/service doesn't bog down but that hasn't helped either. Inserts happen in the background using the elog client app (about 2 or 3 inserts per batch at sporadic times).
|
|
|
|
|
Elog System Requirements, posted by Alan Grant on Fri Sep 15 00:56:38 2017
|
In response to an elog-hang issue I've been having on the Windows platform, I am building a new Unbuntu 14 TLS VM machine to host the identical configuration so that I can more easily debug when the hang happens again. I don't mind beefing up the hardware resources to either eliminiate that as a factor or resolve the problem. I'll have a higher end CPU to deal with 20 to 50 clients doing searches through the data (since the elog configuration currently does not provide a setting to limit how far back it can search with Quick Filters - pretty please add this basic setting!), but the main question I have now is what is a good amount of memory to add to the VM? I suspect even with 30 concurrent searches going CPU power will have more impact than memory in the case of elog. Can someone please confirm my suspicion and also recommend a suitable amount of memory I should install? My data volume is about 25 MB, all textual (no attachmemts), and the number of daily files goes back about 5 years. Any other tips for the build is very welcome. |
Drop down order, posted by Alan Grant on Mon Jan 15 16:46:58 2018
|
Can you please make a change to have the Quick Filter ComboBox controls reference the field type for ordering purposes?
For example, if attribute Lot Number is Type Numeric then the ComboBox should be listed in numerical order instead of alphanumeric (eg: 1,2,3,11,21 vs 1,11,2,21,3).
Best Regards. |
|
Re: Unwanted double entries eg. double clicking submit button, posted by Alan Grant on Wed Feb 20 21:56:32 2019
|
I'm also happy to see this change implemented as we've had to deal with the same issue at times as well. Will this change be incorporated into the latest version (314-2, aka elog-latest.exe), or will there be a new version release (that is not in Changelog yet)? If so, can you give any ETA on this new code availability?
Also I noticed that the Elog Home page still says "Current version is: 3.1.2". I assume that only means it hasn't been updated, not that it means it's the current STABLE version and subsequent releases are beta -- please correct me if I'm wrong. I just want to make sure I understand how the versions and releases work.
Endless thanks for this product and all your work Stefan.
| Stefan Ritt wrote: |
|
I just committed some code which disables the "Submit" button after the first click and replaces the text with "Please wait...". So double submits should not be possible any more.
| David Pilgram wrote: |
|
I too have this as an occasional issue, although in my case due to a dodgy pointer. I too manually delete the entries.
Interestingly, it gives double entries - and thus the start of a branch - even in logbooks were branches are not allowed.
| Finn Junker wrote: |
|
I'm having a minor issue that were getting double entries due to the user is using the "submit" button more than once.
I seems like when there is a lag either on the machine or on the network it is possible to tap the "submit" button more than once resulting i a double or triple entry containing the same text and a almost identical timestamp.
Is there a way to aviod this?, my "solution" so far has been to select the entries and manually delete them. I'm using Elog version 3.14
Kind Regards Finn
|
|
|
|
New feature request for Options list, posted by Alan Grant on Wed Feb 20 22:24:05 2019
|
Is it possible to include an option in the next release to have the Options list reference a text file of attributes rather than explicity listing the attributes in the Config file directly?
This would make it much easier to maintain a particular list that is referenced in several log books. |
|
Hiding Logbook tabs, posted by Alan Grant on Thu Mar 28 21:55:14 2019
|
I know you can restrict access to logbooks on per user basis but can anyone tell me if it is possible to HIDE certain logbook tabs/groups on per user basis (ie, so they would just see logbooks they are authorized to see)? |
|
Last default = <n>, posted by Alan Grant on Mon Apr 29 02:41:15 2019
|
According to the doc on the setting "Last default = <n>" it is intended to restrict Quick Filter searches to the last (n) days of entries, however when I add "Last default = 28" to config it still returns entries far older then 28 days when using the filter search. Am I misunderstanding, or implementing it wrong? We must restrict searches through QF to reduce load.
A couple other side questions which are related:
The doc on this setting says: "Last default = <n>
Some logbooks are very big ans searching through all entries with quick filter can be time consuming. This option sets a default value for the Date quick filter, so that by default only the <n> days are displayed." ...... I'm assuming that by saying "only the <n> days are displayed" that it has actually only searched <n> days ................ vs seaching ALL the logbooks but only displaying <n> days. Please confirm.
Also, why wouldn't the "28" days spec display in the Select Period box beside the Quick Filters? Otherwise it appears to me to be a separate and conflicting item. Best regards! |
|
Re: Last default = <n>, posted by Alan Grant on Mon Apr 29 17:48:13 2019
|
I just changed to Last default = 31 and even restarted elogd.exe but it still finds a given string older than 31 days. Also the default Select Period box value next to the QF's remains at All Entries. My understanding is it shoud behave like Find, where Show Last Default = 31 should default the display to Month, and also only search within the last month of entries.
| Stefan Ritt wrote: |
|
Indeed the description of the "Last default = <n>" option is not clear. I just updated that. If you have a quick filter "Date", then you have different options there like "Last day", "Last 3 days", "Last week", "Last month", "Last 3 months", "Last 6 months", "Last year". Each of these options has an underlying number of days, which are 1, 3, 7, 31, 92, 182, 364. In the "Last default = <n>" option you can only choose one of these seven options. Then this option is selected by default and the filter is applied accordingly. This lets you still change the date filter to a longer time period for example, but then with the penalty of longer search times.
So what you need is "Last default = 31" instead of 28 and it should work.
Stefan
| Alan Grant wrote: |
|
According to the doc on the setting "Last default = <n>" it is intended to restrict Quick Filter searches to the last (n) days of entries, however when I add "Last default = 28" to config it still returns entries far older then 28 days when using the filter search. Am I misunderstanding, or implementing it wrong? We must restrict searches through QF to reduce load.
A couple other side questions which are related:
The doc on this setting says: "Last default = <n>
Some logbooks are very big ans searching through all entries with quick filter can be time consuming. This option sets a default value for the Date quick filter, so that by default only the <n> days are displayed." ...... I'm assuming that by saying "only the <n> days are displayed" that it has actually only searched <n> days ................ vs seaching ALL the logbooks but only displaying <n> days. Please confirm.
Also, why wouldn't the "28" days spec display in the Select Period box beside the Quick Filters? Otherwise it appears to me to be a separate and conflicting item. Best regards!
|
|
|
|