Webcast: Automated 3rd party package testing - Transcript

company JJB
JJ Bateman|Updated May 12, 2021
image (7)
image (7)

Thu, 3/18 • 35:21

SPEAKERS Lex, JJ, Jake, Jordan, Kelly

Lex  00:00 ---  Hey, everybody, it's Lex from PDQ.com. We're gonna take you through automating your testing. Now, there's a debate, do you test Do you not test? Obviously, you see me do stuff I test live in production, but it's always definitely not the best way to go. I know this subject we covered before. But you know, surprisingly, we had an epiphany on better ways to do it. So let's jump in. First of all, Jordan, question for you.

Jordan  00:28 ---  All right.

Lex  00:30 ---  Why would you test man?

Jordan  00:32 ---  Well, I'm a big advocate of testing, because in my younger days, I did not really, yeah. So at one point, we're putting in new load balancers. And they had a different account for login as a service. And I manually put that into the domain controller instead of that group policy to it and overwrote all the existing log on as a service accounts. Except for the one for the load balancer.

Lex  00:55 ---  You had a bad day. 

Jordan  00:56 ---  Oh, that was a very bad day. Okay, but it took like an hour or so for the consequences to really reach us. So I didn't even spend time together. It was I brought down our entire environment for I'd say a day. Wow.

Lex  01:10 ---  Good. So not testing can be expensive.

Jordan  01:12 ---  Again, yeah.

Lex  01:13 ---  Alright. Well, let's do this, I'm going to run through building a schedule, just the production schedule, and I'm gonna show you the old way of building a test schedule that's automated. And then we'll show you guys the new way we came up with, we're gonna roll with Chrome again, this is nothing new building a schedule. So we're gonna do a new schedule. We'll call it Chrome. Okay, triggers, we'll run this, let's just be consistent. I like to run stuff on Friday after hours. So we'll say we'll go up to 8pm. Hey, targets, we're gonna let inventory select those targets for us. You know, why not reduce the amount of work we have to do right? With that, when you're selecting your targets, obviously, you know, Chrome old wood, in this case, what we're going to be looking for is the chrome old collection. Okay. verifier package, again, options stopped point targets once they succeed, and we give this an okay. Now, the story with this is what's gonna happen on Friday? It's gonna deploy Chrome to all the machines that have old Chrome. Now, do you want I mean, again, testing this, we need to test this before, you know, Friday to make this test happen, or to do that, a couple of things we need to discuss for so with that. Jumping into preferences, you'll get auto downloads right now, I've got my auto download set to 14 days. We did that so that I had something to show you guys with auto downloads today. But normally, that would be seven days, right? In my case,

Jordan  02:51 ---  I think that's the default on the product.

Lex  02:52 ---  It is and it's probably a pretty good rule of thumb, right? I don't like testing for people. So let's let everybody help test hands, right, if there's a problem. With that, we're going to need to get a version down earlier, so that we can actually do some testing. To do that, we're gonna go to the package library, and we're going to go download another version of Chrome. And we're going to change that auto download the great thing, again, the global setting is set in preferences, and you can add a per package level, change the auto download option. So as soon as that's done with the check there, and open that up and go to options, okay, we're going to set this to immediate, so as soon as one is available, it's going to download it. And it's probably a good idea to indicate that this is a test version of Chrome. So I'm going to pen-test on there. Any naming conventions you've found useful, I just depended at the end test.

Jordan  03:57 ---  I prepared at the end that way I know. I get anything that's uniform is fine. I wouldn't recommend putting it in the middle. I'm not sure if that would update properly. But beginning and end, 

Lex  04:05 ---  I think you need to keep it at the end. So I'm just gonna put a test on it. So now we can see the test. You'll also notice right here, the approval automatic immediately, okay, as opposed to your settings from preferences, you can now tell it that it has had that changed. Okay, now that we've got the test package, we need to build a collection of machines to test with. Now because we're going to test Chrome, I am going to open up Chrome old. Actually, I'm going to do this first. We'll build a new dynamic collection. We're going to call these test machines right. 

JJ  04:46 ---  Still got that highlighter up.

Lex  04:47 ---   I mean, you know, JJ wanted you to enjoy that highlighter. 

Jordan  04:50 ---  He was really highlighting it.

JJ  04:52 ---  Highlighting something.

Lex  04:54 ---   Something. Oh, sorry. This is going to be a static collection on a dynamic collection.

Jordan  05:00 ---  We're off to a roaring start, boy,

Lex  05:02 ---  wow, okay, test machines. Now, I could go in and learn to type, but we all know that's never gonna happen. I'm going to do this. I'm gonna give that up. Okay, there's test machines. I'm going to go click on Chrome, because I want to use some of these machines as tests. And I'm going to select on Conrad coolcat Elmer Fudd foghorn I don't know, Penelope, speedy. I'm just gonna drag and drop those guys into my test machines. There they are. Okay. Another thing I'm going to do just is I'm going to grab some machines that don't have Chrome on it, because you know, it will make a difference. So we'll go chum clod crony, drag those ins now I've got nine test machines. When you're selecting your test machines, a good rule of thumb. You know, if you've got a myriad of OS'S grab some different OS'S, if you got different groups, a software like accounting has different software than development has different software than, you know, I don't know if the other departments probably want to mix that up a little bit. And then I do like to go pick one that, you know, will just kind of piss somebody off. So JJ, which one microverse one or two 

JJ  06:12 ---  Those are yours.

Lex  06:13 ---  Okay, now, I'm not gonna do that. But so this, these are test machines. And that's where we're going to kind of stick this now, the previous way to build testing is I would build a production, we've got a production collection we're going to run against which is Chrome old. So we need to run, in this case for the Chrome test against machines that are in Chrome hold and part of the test machine. So to do that, what I do is build a new dynamic collection and nest collections inside there. So go to the computer, go to collections is a member of we needed to be a member of Chrome old. Right. And we also needed to be a member of and my clicking isn't there, we have crawled. And we'll also need one that is a member of test machines. So that we're actually testing machines that are upgrading, right. Any, any thoughts on that, Jordan?

Jordan  07:11 ---  But just isn't that what you want to make? Sure it is all for that. So they're a part of both of those collections? If it's the any, then you're gonna get all the Chrome old and all the test machines, it's just the the

Lex  07:23 ---  not gonna not be what you want. Yeah. could have a bad day? We'll give that No. Okay. And so we should see, there you go. There's six machines that are Chrome old, and also part of the test machines. So at this point, I'm going to go to our test, we'll make a new schedule for testing. We'll say let's run that. We'll call this test. Dash Chrome. We'll run that daily. Yeah, test it every day. And we'll do that. Again. We'll do that after hours. We'll do that at 7pm.

JJ  07:58 ---  But you don't work at 7pm.

Lex  08:01 ---  Okay, if it's me, we can stop it at noon. Right? I mean, oops, did I just talk myself out? I wouldn't know I work. I work late. You don't know that I really work.

Jordan  08:10 ---  I said, I wouldn't call it we do work.

Lex  08:13 ---  Yeah, that's a good point. We'll stop playing today. And again, for choosing targets. Now I'm going to go grab that new collection we built which is test Chrome, you can hear my mouse wheel scrolling, boom, package, verify that we actually selected it there it is the test one stop pointing to targets once they succeed. give it an okay. And so now I've got my Chrome, which is production and my test Chrome, and how that's gonna work downloads immediately that night, it's gonna run and do your test. This is the original way that I was showing people how to do this. Not bad.

Jordan  08:52 ---  So there is a caveat. Yeah, this one, where we have the 14 day which doesn't set it to 14 days. Yeah, but then the schedule also only runs on Friday. So if Chrome updated on a Tuesday, like it normally does, 14 days later is going to approve, and it's going to be another three days before the schedule runs. A lot of products, you know, generally when they get an update, and you can kind of build around that. But if you only have it once a week, you might have a longer wait period than expecting to patch your systems. So it's just something to be aware of.

Lex  09:19 ---  I think it's a good thing, though. I do because it's like let's say it comes on Tuesday, the next Tuesday, downloads Friday, if it came out on Saturday, Saturday, so yeah, it'd be a little bit longer, but I still think it's okay. 10

Jordan  09:31 ---  days is a long time to not patch for security. anymore. Everybody was Oh no, I'm just saying. If you want it to be less, that's just something to be aware of. It could make it longer than you're expecting.

Lex  09:45 ---  Yeah. Okay. All right, between the debate here, we're gonna take a question.

Kelly  09:49 ---  All right. All right. Here we go. This may be a secret recipe question for you guys. But I'm curious what the usual timeline is for When a manufacturer like Adobe releases an update to when the package library is updated, thanks, as always, Timothy.

Lex  10:10 ---  So go ahead,

Jordan  10:12 ---  oh, well, we have some packages that are considered critical that we have to do that. As soon as those come in, we have people that are working and testing to get those up as soon as possible. I can't give an exact time. But as long as time to test to make sure the new package works, and there's no noticeable errors out the gate. For some other ones. Like if it comes on Patch Tuesday, and it's something that's not widely used, it might take a little bit longer to get that one because they're gonna get the Windows Updates first. Yeah. But we see those pretty quickly when they come in. And it just the time wasted is not time wasted. The time spent is more on the testing and building of the new package than it is not seeing it

Lex  10:47 ---   rule of thumb. Usually within a day we try to be I,I know we try to be when I was doing it, we're trying to be first out to market so it came out. I wanted to make sure the patch was available as soon as we could. I think that's still the philosophy. Yes,

Jordan  11:00 ---  as soon as possible is definitely the overall goal. And we got a really good team on it. So

Lex  11:04 ---  absolutely. All right. Chrome Chrome test.

Jordan  11:12 ---  Are you lost?

Lex  11:14 ---  I feel lost. Alright, so again, Jordan brings up your triggers and you just your timing. Okay. Good to know that. Okay. Now, we found a better way, a better way to do this. Okay. And here's the thought process. And like, you know, doing it this way, I've got to set up a schedule for every single test. And I've got to set up a test, Firefox, test seven zip to collection for every single one. And so I'm talking to Jordan, like, wouldn't it be great if we just do, you know, one scheduled one collection to rule them all. And so we batted around ways to do it in PowerShell. And then, all of a sudden, I'm like, hey, what if we just did this? And it was so simple. I mean, it's,

Jordan  12:01 ---  there's actually a comment in the chat. It says, Let me guess PowerShell which was our first instinct. We didn't need it. Much. I never thought I'd say those words. Wow.

Lex  12:09 ---  Well, that's where we went. So I'm going to show you

Kelly  12:12 ---  Oh, you did? Did Did you seriously just say you didn't need I'm not

Jordan  12:15 ---  saying I couldn't do it with PowerShell. I'm just saying it was unnecessary complication for this very specific task. Okay.

Kelly  12:22 ---  All right. 

Lex  12:24 ---  What did come out of it, though, is we found we could reference other variables and variables, which were going to be shown in another webcast. I don't know

Jordan  12:31 ---  what you can do with it yet, but it feels like you could be awesome. You can do something awesome. We just don't know what yet. Yeah.

Lex  12:36 ---  Alright. Enough build up enough anticipation. Should we take another question?

Kelly  12:45 ---  Okay. Hey, guys, is there a way that we can find the signup parameter for a.xc program, for instance, Adobe pro version, which is kind of a big package. Sincerely, vignesh 64?

Lex  12:58 ---  Okay, um, two places I would go. Sometimes you can run the EFC with a front slash question mark. And you'll get the possible parameters that you can use the first place I check. Okay, that's the second place I check. First place I check is Google Fu. And again, you know, I'm looking for the vendor. And if not, from the better you start, you know, expanding out from that community and that Jordan, your your thoughts on

Jordan  13:23 ---  so I can't verify this is 100% is just what I've seen. If the Exe is just the un INS  then three digit code normally 001 forward slash silent uninstaller. All capitals have worked every time in that scenario for me, but that is a personal observation, not a guarantee. But if it's the specific named executable, I usually work with the same parameter.

Lex  13:51 ---  Okay, let's, let's show them the easy way to do it. Alright. So first thing, I'm going to show you this. Okay, we're going to go in and you'll notice I've got two packages that need to be approved. I've got Adobe Reader, DC, and Windows. We're going to go in here and I'm going to call this. I'm going to rename this to say this is my test package. Okay. And what I'm going to do is I'm going to say this package is a test package under conditions that can only run on machines that are a member of reader DC old. Apparently, typing does help. P FM's sometimes

Jordan  14:35 ---  there was an early recommendation that we get you Mavis beacon teaches typing,

Lex  14:40 ---  they've already done it may happen, it ain't happening.

Jake  14:44 ---  She can't teach his fingers how to be smaller.

Lex  14:47 ---  Alright, DC, let's assume I got the right one here in DC. Okay, your DC your DC looks like Okay, so this is this test is only going to run on DC old. So remember when I had to merge those two nested packages. I don't have to do that now. Okay, I'm gonna save that now just to verify that this actually retains that, we're gonna go approve the, the approval and we'll show you that it actually retains that approval now. It's gonna download and it will save that parameter. And it's so cool. I gotta do a little thumbs. This is cool, but this is super cool. Okay, that that's it, that's okay, that's not gonna abuse me.

Jordan  15:27 ---  I'm not gonna, you know what, I think you're doing a great job. Thank you for that, please

Lex  15:32 ---  do not do that. I don't know how to handle compliments. It's really sweet. I need to drink. So open this up. There it is, you notice it is now the new version, and it retains the collection reader DC old. So I'm gonna do that with our Chrome test. Okay, and we're gonna go in again, we're gonna say this only runs on members of Chrome Old. Okay. So now probably should save that. Yeah, probably the best. Yeah. So now we're going to do this, we're going to do a new schedule. And we're going to call this the new super up, you know, I'm not going to do a new super Chrome test. You know, I should not have tried to make something funny or have to type. I should have just stuck with the new Chrome test. Okay. Now, again, triggers, we're gonna run this daily. And we'll do it. When I get off work, two o'clock. Just kidding. Let's do it at eight targets. All I have to do now is just go and choose the test machines. Okay. So now the package itself is only going to run specifically on machines that are Chrome old, and I'm only throwing test machines at this point. Okay, there's my test, stop deploying targets, once they succeed, given an okay, I'm going to kick this off. Before we do that, let's just go verify, right? So we should be seeing six machines right there nine machines here, but it should only update six of these, right? To test that, we're gonna right click, we're going to say start schedule our test packages. And let's go look at that deployment. Okay, so there's my deployment. That one failed crony not run due to condition. Run due to condition. So we got 123456. Ah,

Jordan  17:35 ---  this does condense your schedules. So you don't need so many of those. But it does, it condenses my collections, your questions? Yeah, it does, however, come with a complication like on a Patch Tuesday, this is going to be sending out a lot of installs. And you might be hearing from the network team.

Lex  17:51 ---  Let's Let's not talk about the downsides.

Jordan  17:54 ---  Well, yet. Well, we can't say here pretend like everything's perfect. We can't Okay, then then there are no downsides. But without there being downsides. Are there any options people have where they can have deployments that aren't going to fail faster to move on quicker, I suppose to make it just a faster process.

Lex  18:17 ---  I don't know what you thought you asked the question. I'm gonna have to say at this point. It's running. Yes. The test machines. This is my test machine, and it's only updating the ones that are old. So it is going to fail. Oh, yeah. You know, we can now think about it. Let's go look at that schedule. schedules. Super cool. Chrome test. Offline settings. Your settings from packages. Yeah, offline settings vary. So our offline settings in preferences, this does make a difference. So under performance, deployments, offline settings, here we go. So I would highly suggest using the right key to turn the highlighter on ping before deployment. Okay, so for machines off, boom, it's not going to try it's just going to move right on. What are your thoughts on this?

Jordan  19:11 ---  That's kind of the big one, that the Send wake on lan, we'll make it a little bit longer again. But that way, if they're just asleep, it can be woken up? It will, you'll get more machines that way. So it might not be a bad idea to have, that's going to kind of remove the speed

Lex  19:24 ---  definitely if I was going to do the wake online, I wouldn't do it as a global setting. I would do it at the schedule or the package level specifically, just so that you're not hitting everything. 

Jordan  19:35 ---  And as Colby pointed out as far as member of collection condition, that one fails really quick on its own. Yeah. So those ones will go fast no matter what.

Lex  19:45 ---  So in this case, I would set them right here. So just the test machines are trying to be woken up as opposed to every machine you might send it to. And you can also do that at that point at level. So good, good point. So in this case, I can now instead of just calling this the super cool Chrome test, I can take out Chrome. And because each of these packages are only going to install on the machines that are in test and whatever package that's old, I can now add additional packages to this. So that one, if I add it, and here we go, there we go. Three times, and then I got it. There we go. Okay, so now I can test all of these with one. downsides to this?

Jordan  20:39 ---  For multiples? I mean, it depends on what you're putting in there. Like if you're doing the Well, I mean, Windows Updates gonna be a one per machine either way. But if you have, say, six or seven packages, and there's over a longer install, you might need to look at your total deployment window to make sure it doesn't close out prevents finished, if all five of those have to install at once back to back, or is that does that work per package? Or is that just all considered one deployment and

Lex  21:04 ---  for the timing? When it kicks off, it's going to look to see if any of these I mean, it's going to go look at each one of these. So if there's nothing new, or nobody is in, like Adobe Reader, old, oh, no, should we kick it off?

Jordan  21:19 ---  Let's give it a go, shall we?

Lex  21:20 ---  Alright, let's see, we've already updated Chrome. So now, I didn't save that. Well, while I'm fixing this should we take

Jordan  21:29 ---  we fix that? Someone in the chat said, you have to be careful with using collections as it conditions, because it's good as the latest inventory scan. In this case, we're looking at the latest old software. So if it updates as either the software is now uninstalled, you don't want to do it anyway. Or the software is already updated, it's going to skip. So no matter what the scan is, if it updates, it means the software is already updated on that one. 

Lex  21:52 ---  Okay, so our new super test, don't let's take a question and I'll run this we'll find out what happens.

Kelly  21:58 ---  Here, Lex and Jordan, I need to run a heartbeat job as a priority on a collection. There are other existing heartbeat jobs. This is a limited time job that has an end date that needs to be run First. The next time the computer connects, is there a way to set a heartbeat to automatically run as a priority? Sincerely, Russell? Okay.

Lex  22:20 ---  Automatically,

Jordan  22:22 ---  I don't think I think of it as a way to elevate priority once it's already running.

Lex  22:26 ---  Yeah, once it's running, you have to actually that's a physical, you gotta go right click on it and say, you know, prioritize, and it bumps it. So could you? And this is not, should you? But could you PowerShell change that. So okay, things you're gonna run into, right? So you've got one, there's got to be a trigger, the machine came online to run and it's now in queue.

Jordan  22:47 ---  It has to be like a SQL query to find out if it is currently updating. And then if it is, you'd have to then

Lex  22:54 ---  then you'd have to run a heartbeat,

Jordan  22:56 ---  then you have to update the den and do a sequel change to manually set priority, because there's no command line to do that? Yeah, I would say don't don't,

Lex  23:05 ---  I think you don't mean, can you? Maybe it's a lot of moving parts there. But probably I would be careful of causing problems to your database. So Alright, let's kick this off and see how many I wonder how many of these are actually in Adobe Reader. Oh, let's go. At least at once we can see one of these kickoff right, reader.

Jordan  23:28 ---  I'd be surprised if any of these machines had it.

Lex  23:33 ---  11 reader, old old 17

Jordan  23:38 ---  what do I know?

Lex  23:40 ---  Right. I think they're in there. Let's go find out. Conrad. All right. Yeah, we'll kick it off. Should we at least see one right? All right, right click. Let's see what happens here. Start schedule. All packages. So this thing runs tonight at 8pm kicks them all off. We've got not run due to condition not run due to condition or new to condition

Jordan  24:10 ---  because a lot of those you already ran the Chrome one. So there is not going to be any Chrome old.

Lex  24:15 ---  There you go. Speed Penelope Connor, there's Conrad. Let's go take a look at the deployment is doing or Adobe Reader. So it's gonna make the attempt on all of these that fall in that category. So there you go.

JJ  24:29 ---   I can't believe how uncentered that is. 

Lex  24:32 ---  What we mean, dead center. 

JJ  24:34 ---  Now. That was so far off. 

Lex  24:38 ---  Okay, I want to throw another thought out. Okay. So if I add all my packages to testing to the one, right, so I got one, it runs against my test machines. I'm feeling pretty good. I'm gonna get reasonably good tests. If I get flooded with a bunch of new packages at the same time it's a Patch Tuesday. Would I be better off to stagger that, let's say have a Monday and then I can pick them packages, they could test it on Monday. You know me to do it that way. Right? I would do, again, a very similar style thing to start with Chrome test. Right. And we'll make a new schedule. We'll call this the smart. Monday test, right? Smart five, apparently I'm not smart. But that's nothing new, smart Monday test, right. And again, triggers for that one would be we would run, in this case weekly on Monday, right? Boom. And, again, then I just add the ones I want to run on Monday. I mean, not a bad thing. And the nice thing is, I still have one collection, and then I can divide up. What gets tested Monday, work is tested Tuesday. What are your thoughts?

Jordan  25:50 ---  I don't like it for tests, maybe for production. But for tests, I mean, it sure it might mostly update one time. But if you have an out of band one, this is something seriously wrong. You want to test that. So for the test, I don't want a test run and check every day no matter what. Okay, that's because I want the test to always be up to date with the latest one. If you're looking for production to expand it out that way, that's probably not a bad idea. But at that point, if you do get an out of band, like, let's say Windows update is causing a blue screen, we did network printing.

Lex  26:19 ---  Okay, so then instead of this, how about I do all my third party testing in super cool test? And I got my super cool Windows Update. And I could run those specifically on a different day. What do you think? Is that still we run into the out of band

Jordan  26:37 ---  Lukas that we just actually had one new Windows update that came out two days ago, okay, because of an update that was causing blue screens.  So it's one of those where for the test, I'd want the test to go no matter what all the time every day for the production, you could probably try to work around it. But in that case,

Lex  26:55 ---   it is scheduled for production anyway. So yes, it would kind of moot point, wouldn't it? What do you guys think I'm just wondering, because this is something I thought of this morning in the shower, I'm like, Huh, I could do Mondays, I can do Tuesdays and you know, it's one, you know, one schedule, looking at a collection for tests. So

Jordan  27:12 ---  your test environment by default is going to be a smaller environment anyway. So being more aggressive, I don't think it will be a problem. And I just I would always want the test on there as soon as possible. Because you want that thoroughly tested before you go before you go live. And then for the other one, if you are only going to have it where you're specifying like only after the second Tuesday, or like the week after that. So you give it a week to test or whatever. Then if you do have an out of band, you want to keep an eye on that you might want to manually kick off the schedule. That's extra work. It is extra work, but I'd rather do that extra work than wait a month for it to loop back around point.

Lex  27:47 ---  Point. All right, should we take some questions? I think I'm gonna have to agree with you, man. I'm testing more aggressive schedules. Probably better to do questions. 

JJ  27:58 ---  Yeah, let's do a speed round. 

Kelly  27:59 ---  Yeah, let's go speed round on this one. Hey, guys, some of our workstations upgraded to Windows 10 2004 2009. Those auto download packages were not set to manual the way 1909 and earlier is, is there a way for me to un-approve the march 9 update and put it back to waiting for manual approval. We're usually one month behind. Thanks, Josh. See,

Lex  28:22 ---  you leave that up for a second. There's a way to un-approve the march 9 update. Jordan, thoughts on approving

Jordan  28:33 ---  I mean, we're a month behind. If you see that coming, I could see maybe disabling the schedule for a short time and that would still approve it. But it's not going to deploy out and then when the next one comes out, it will skip over that.

Lex  28:44 ---  Yeah, I was gonna say if there's a problem, the first thing you do is you go into schedules, and we can show us right click on it and just disable the schedule little green guy goes gray, and there's no next runtime. And then after the problems taken care of reenabled that scared us enough to delete schedule and don't have to spend a lot of time approving and un-approving So yeah, I would I would definitely just disable schedule in that case. I did it we didn't know we didn't Did we?

Kelly  29:17 ---  Alright, ready for another one, guys. Hey, guys, thanks for putting on the webcast. I would like to use one drive to store my packages and install directory is one drive supported. Lorenzo b

Lex  29:31 ---  alright, UNC paths. Alright, so can you UNC path to OneDrive?

Jordan  29:36 ---  I don't know if the credentials are different. But I imagine you can as long as it has a UNC path and the credentials that you're deploying or can read can get to it. I don't think it should be a problem. But that's speculation. I haven't tested with OneDrive.

Lex  29:49 ---  Yeah. I think that would probably be the case. So

Jordan  29:53 ---  Oh, no, he says no UNC path only HTTPS path.

Lex  29:58 ---  Hmm. I've never tried

Jordan  30:01 ---  Yeah, I gotta say, probably not, but I haven't missed it. I've only done UNC pass. Agreed? So no.

Kelly  30:10 ---  Okay. That was a long way around to know what we learned as we went, Hey, guys, is there a way to reduce software download footprint by combining the location of the files Chrome Once approved could just use the downloaded file from Chrome test? Eric B.

Lex  30:30 ---  I think the issue you'd run into there is obviously timing and immediate versus so I don't know the way timing works. And that, I would have to say, I don't think there's a way easily to do that.

Jordan  30:40 ---  You need to have this, the second schedule for production has to kick off based off the test schedule running. That's right. So you might be able to do like a variable, the updates when that schedule kicks off, and then it adds that variable moves over the production, which then runs X number of days after.

Lex  31:00 ---  But that would be a fun one to make for a web show. But

Jordan  31:02 ---  there's a lot of variables on that one, where if it kicks off on a Wednesday, like how do you specify it's definitely three or four days later?

Lex  31:09 ---  It's a lot of moving parts.

Jordan  31:11 ---  There's a lot I can't wrap my head around all of that. It's not easy in the minute timeline.

Lex  31:15 ---  Maybe we'll have a solid maybe. I like the way you think, though, because it looks like you're trying to solve it. So.

Kelly  31:24 ---  Alright, well, so far, we've had a no one to maybe. Let's see what we get. Hey, guys. Have you seen pester tests? Or how have you seen pester testing used in PDQ packages? I've used it in the past to validate a computer's configuration and send an email to a team's channel with the results. What are some interesting use cases you've seen? Sincerely, Andrew

Lex  31:47 ---  testing for a while watch it.

Jordan  31:49 ---  I was we I don't we were just starting to look at pester when it is pester? I don't know. It's a PowerShell module that is built in testing for cmdlets. or not, or for your script, not for commands before your script that will basically look for common errors you basically supply expected results with pester, and if you're not getting that when you run the command. Oh, so it's basically testing to make sure your script works before you're in production. What

Lex  32:15 ---  is something that we do all the time? It is you and me.

Jordan  32:18 ---  pester is awesome. Okay, it is absolutely awesome. So I don't know if they're using or not, because I'm not in the process. But I'm sure they are doing some sort of testing everywhere. They need to have it. Yeah.

Kelly  32:33 ---  All right, two more questions. Boys and girls. Have you guys ever thought of incorporating news for related ms patch with available download IE? no error slash crashing? David G.

Lex  32:48 ---  News. Good, Jordan.

Jordan  32:49 ---  So when something comes out, like the windows one of the blue screen, we generally put out a blog on that.

JJ  32:56 ---   If you haven't subscribed to our blog, do that now. Yeah, because we were really consistent about both our Patch Tuesday updates and our out of band ones when those happened.

Jordan  33:08 ---  Yeah, so there's nothing in the product itself, although it could that be added the banner, I don't want to miss Oh,

Lex  33:15 ---  no, there's some information that comes with it. The description. I mean, I guess there's that. And then again, if you noticed, when I selected it, there is a smooshed up box that has that in it. So you can attempt to read it against the package library right here. So

JJ  33:32 ---  the blogs are a lot more reader friendly.

Lex  33:36 ---  But you're gonna make it friendly for people. Yeah, the blog, guys. So

Jordan  33:39 ---  yeah. So for the recent windows one brought to that blog, and he has in there PowerShell that could remove that KB if you wanted to roll back instead of wait for the patch, as well as going over more information about it.

Kelly  33:54 ---  Okay.

Lex  33:56 ---  Find out the answer is yes. The

Kelly  33:57 ---  answer. Yeah. Yes. Yeah. Final question of the day. Having the package in the cloud helps with users working remotely any plans to include cloud storage to deploy apps? Lorenzo b

Lex  34:10 ---  I'm gonna go with Yeah, yeah, there are.

JJ  34:13 ---  Another Yes.

Lex  34:14 ---  Oh, hell yes. Is that what I was supposed to say? No. Yes. Yeah, it's definitely we're definitely looking into that. It's, I guess, the way of the future. Yep. So anything else?

Kelly  34:27 ---  That's it?

Lex  34:28 ---  I gotta come up with a closing statement now. Dude, as I say not to test guys before you have a really bad day. Have a kind of a bad day out a really bad day. Testing is important, guys. Thanks for watching. I'm election PDQ.com.

Kelly  34:48 ---  Thanks for joining our webcast today. Congratulations David G and Timothy m winners of PDQ swag. Send us your info at the webcast@PDQ.com. If you've got additional questions. Feel free Reach out to our support folks. Thanks again for joining us, and we'll see you back here next week.


company JJB
JJ Bateman

JJ is a technical creative. He finds joy in programming, automation, and in participating in the artistic sides of things. You'll often find him drinking on Thursdays on the PDQ webcast, rambling and raging remotely at the Bermuda Triangle/heisenbug-tier tech at the studio in SLC.

Related Articles