I've been thinking of doing the same thing. Try out NearlyFreeSpeech.net. They are awsome. You get ssh and ftp access and only pay for what you use. I think my domains are about $8 / year, hosting costs me about $1 a month on a busy month. This is more then $3 / year, but it's a really good price IMHO and scales easily. I could easily decrease my costs by turning off my email forwarding and deleting some stuff, their base price is 1 penny / MB stored and $1 per GB transfered each month.
I'm posting this reply since I found your information useful. It turns out our SAN had 97% of it's disk space allocated to LUN's. These LUN's were nowhere near full, but the SAN still doesn't like this. Xiotech indicated their SAN's start to degrade substantially after they become 80% allocated. We added a bay of disks and have seen a vast improvement in performace. Since our license doesn't allow us to use all the space we added we shouldn't have this problem again.
Thanks for your insight.
I'm posting this reply since I found your information useful. It turns out our SAN had 97% of it's disk space allocated to LUN's. These LUN's were nowhere near full, but the SAN still doesn't like this. Xiotech indicated their SAN's start to degrade substantially after they become 80% allocated. We added a bay of disks and have seen a vast improvement in performace. Since our license doesn't allow us to use all the space we added we shouldn't have this problem again.
The metaphor Jesus used was more about the Shepherd then the sheep. Because sheep are so stupid Shepherds are known to go to great lengths to care for them. I've never heard anybody preach that we should be sheep, only that we should care for others like a Shepherd cares for his sheep. It's hard not to assume your misinterpretation is malicious or deliberate, but I realize your level of discomfort with this topic probably makes it difficult for you to understand so I will give you the benefit of the doubt.
A couple people have mentioned moving my swap volume to a local disk. I will do that because it will hopefully help, but there is also no reason for it to be on the SAN, it doesn't facilitate Disaster Recovery. I should have thought of that sooner.
I should have mentioned our roaming profiles are only used for Citrix. We don't publish a desktop. I have noticed that some users who have excessive login times have a profile that uses much more space on disk then the size of the profile. Recreating the user's profile complete seems to clear this up. That's one of the things that makes me look at disk fragmentation.
I don't have any good way to measure throughput on my fibre cards or McData switch (this I should probably start monitory). The bottleneck seems to show up in the disk load. I attached some graphs here to show what I'm looking at.
Our page file is on a dedicated partition on the SAN also. I do notice that it is usually 80% utilized and at night when our backups run it goes close to 100%. Our diskload also spikes at that time, but not as high as it does in the morning. When I get the high diskload spikes in the morning the page file seems to maintain at 80%. I speculate this is because the things being accessed in the morning are more random and the server doesn't load them into memory or use the page file. While the backup client probably stages things to memory and uses the page file.
I don't think this is a problem. Our roaming profiles are only used for Citrix, which is where the majority of our users are. I have user directory going to a network drive (on the same server as profiles). I've never seen a significant hit on the network cards on either the file/print/pdc or the citrix servers. I've also never seen significant hit on any of switch ports (servers are on gigabit).
This is probably a huge part of our problem. The SAN was sold to my predecessors as a kind of "magic box". I've avoided taking any training on it because it seems so simple to administer. In hindsight this was a mistake and I should probably take the effort to learn more about how the physical architecture relates to the assigned disks.
I use Big Sister to monitor all my servers. I get nice graphs that show memory, CPU, network load, disk utilization, etc. I looked and looked at this trying to find the cause of my problems. People complained about slow login times, sometimes they would get temporary profiles because their roaming profile would time out. The also complained about slow access times in our SQL dependant EMR (Electronic Medical Records) system. All my graphs showed everything within an acceptable range.
I finally found an SNMP query for "disk load". This purports to be a percentage, but I've seen it showing way over 100, sometimes as high as four or five hundred. If it gets above 50 or 60 people start to complain. My disk load spikes in the morning when people are logging in, it generally goes to about 80% or higher on my graphs.
My SQL server doesn't have these problems and I have yet to find a suitable way of monitoring the SQL log where I think the problem is originating.
I've talked to three people who are either certified on this product or work for the manufacturer. I got three different answers with little in the way of explanation.
Actually, since my software is a perfect duplicate of their software, made by me on my machine, how can they own it? I owned the blank CD, and all the equipment reading the "CD they own".
Everyone's talking about Certifications and Degrees when what he wants is a job. I recommend really polishing up your resume, get several people to look at it and pick it apart for you. It should be no more then 2 pages and preferably one page. I like to stick in a skills section for jobs that might be data mining for key words.
Step two is to get your resume to as many head hunters as possible, this can bite you in the ass if you are applying to a job that has already received your resume from a contract company, but in this entry level position it is an acceptable risk. Head hunters are good because once you add their embellishment to your embellishment you look damn good. They also have short term jobs that allow you to gain experience quickly.
Here is the most important part. Customize your cover letter for every job you apply to, contact each company you apply to no more then one week after they should have received your resume. Try to speak to a hiring manager and verify they have received your resume and find out when they are interviewing. Ask for an interview, this gives you a chance to clear up any concerns they have with your resume, maybe you asked for too much and can afford to take less, maybe they wonder why you are so cheap (agree to take more). After your interview keep in touch, send a thank you letter. If you are skilled but inexperienced you are really asking them to take a chance on you. Convince them you are worth it.
I beg to differ with you. I work for a Mental Health company with about 250 employees. We have one site with about 100 users, 5 sites w/ 20+/- users, and over a dozen sites with less then 3 users. We have a three person IT department.
There is no way this would work without thin client. Most of our locations use either a direct T1 or residential internet service w/ an IPsec VPN. Some of our sites need computers to handle a dynamic VPN. These sites are way more trouble then the sites with thin client terminals. Our machines are configured to be easily switched out with no data stored locally, but administering windows locally is still a pain.
I have tried linux as the thin client OS, but VPNC only runs under root, I've managed to get it to connect w/ scripts and sudo, but it's unstable.
That's actually an example of the doctor trying to screw you and the insurance company stepping in and saying, "Step-off, this is my bitch!". It pisses me off to see the statements I get where the insurance company indicates what they were charged and what they paid. My wife just had some tests that were I think around $112 & $70. Insurance paid around $12 & $17.
I haven't seen it here yet, but I use NearlyFreeSpeech.net. They now offer their own domain registration service. From their ideals listed on the page. I can't imagine them caving to any bogus threats. Their prices and support are also great!
I use JIRA. It's nice and very customizable. I had to change some things since it is very software centric and my issues are network and hardware related. It is good for tracking entire projects and individual trouble tickets.
If I could add one thing it would be the ability to auto-open and close a simple ticket for things like password resets and printer installs.
I beg to differ. I'm in a department of three, director, admin (me), & tech. We support about 200 users at 17 different sites. Only one site has more then 30 users. and more then half only have 2 users. If we didn't use thin client we would need at least double, if not triple the staff. We do use some full PC's, mostly laptops. They are used like a terminal to access citrix, nothing is stored locally.
Some things need work, but all in all it's a remarkably cost effective solution.
I had a similar situation with a prospective employer several years ago. It was a contract company (starts with a k and rhymes w/ dforce). The recruiter offered me a benefits package which turned out to be their internal employee benefit package, not their contract worker benefit package. We went back and forth, they offered to give me the better benefits and cut my hourly rate (no way!).
Finally they agreed to honor the original offer and I had them send me the paperwork. I sat on the paperwork until I heard from another company I was waiting on. The gave me a better offer and I called the orginal recruiter and politily declined a couple days before I was scheduled to start.
I'm probably blacklisted, but realistically I was afraid they would replace me immedietly since I was obviously costing them more then the usual going rate.
Slashdot Mirror
I've been thinking of doing the same thing. Try out NearlyFreeSpeech.net. They are awsome. You get ssh and ftp access and only pay for what you use. I think my domains are about $8 / year, hosting costs me about $1 a month on a busy month. This is more then $3 / year, but it's a really good price IMHO and scales easily. I could easily decrease my costs by turning off my email forwarding and deleting some stuff, their base price is 1 penny / MB stored and $1 per GB transfered each month.
I'm posting this reply since I found your information useful. It turns out our SAN had 97% of it's disk space allocated to LUN's. These LUN's were nowhere near full, but the SAN still doesn't like this. Xiotech indicated their SAN's start to degrade substantially after they become 80% allocated. We added a bay of disks and have seen a vast improvement in performace. Since our license doesn't allow us to use all the space we added we shouldn't have this problem again. Thanks for your insight.
I'm posting this reply since I found your information useful. It turns out our SAN had 97% of it's disk space allocated to LUN's. These LUN's were nowhere near full, but the SAN still doesn't like this. Xiotech indicated their SAN's start to degrade substantially after they become 80% allocated. We added a bay of disks and have seen a vast improvement in performace. Since our license doesn't allow us to use all the space we added we shouldn't have this problem again.
I recommend you look at Monowall for a boots from CD OpenBSD firewall router, or I prefer pfsense because it allows you to install to a hardrive and has more features.
The metaphor Jesus used was more about the Shepherd then the sheep. Because sheep are so stupid Shepherds are known to go to great lengths to care for them. I've never heard anybody preach that we should be sheep, only that we should care for others like a Shepherd cares for his sheep. It's hard not to assume your misinterpretation is malicious or deliberate, but I realize your level of discomfort with this topic probably makes it difficult for you to understand so I will give you the benefit of the doubt.
do you mean 90% full of data, or 90% full on allocated LUN's?
A couple people have mentioned moving my swap volume to a local disk. I will do that because it will hopefully help, but there is also no reason for it to be on the SAN, it doesn't facilitate Disaster Recovery. I should have thought of that sooner.
I'll try to post here if that helps or nor.
I should have mentioned our roaming profiles are only used for Citrix. We don't publish a desktop. I have noticed that some users who have excessive login times have a profile that uses much more space on disk then the size of the profile. Recreating the user's profile complete seems to clear this up. That's one of the things that makes me look at disk fragmentation.
I don't have any good way to measure throughput on my fibre cards or McData switch (this I should probably start monitory). The bottleneck seems to show up in the disk load. I attached some graphs here to show what I'm looking at.
Our page file is on a dedicated partition on the SAN also. I do notice that it is usually 80% utilized and at night when our backups run it goes close to 100%. Our diskload also spikes at that time, but not as high as it does in the morning. When I get the high diskload spikes in the morning the page file seems to maintain at 80%.
I speculate this is because the things being accessed in the morning are more random and the server doesn't load them into memory or use the page file. While the backup client probably stages things to memory and uses the page file.
I don't think this is a problem. Our roaming profiles are only used for Citrix, which is where the majority of our users are. I have user directory going to a network drive (on the same server as profiles). I've never seen a significant hit on the network cards on either the file/print/pdc or the citrix servers. I've also never seen significant hit on any of switch ports (servers are on gigabit).
This is probably a huge part of our problem. The SAN was sold to my predecessors as a kind of "magic box". I've avoided taking any training on it because it seems so simple to administer. In hindsight this was a mistake and I should probably take the effort to learn more about how the physical architecture relates to the assigned disks.
I use Big Sister to monitor all my servers. I get nice graphs that show memory, CPU, network load, disk utilization, etc. I looked and looked at this trying to find the cause of my problems. People complained about slow login times, sometimes they would get temporary profiles because their roaming profile would time out. The also complained about slow access times in our SQL dependant EMR (Electronic Medical Records) system. All my graphs showed everything within an acceptable range.
I finally found an SNMP query for "disk load". This purports to be a percentage, but I've seen it showing way over 100, sometimes as high as four or five hundred. If it gets above 50 or 60 people start to complain. My disk load spikes in the morning when people are logging in, it generally goes to about 80% or higher on my graphs. My SQL server doesn't have these problems and I have yet to find a suitable way of monitoring the SQL log where I think the problem is originating.
Xiotech 3d 1000, w/ qlogic fibre channel cards.
I've talked to three people who are either certified on this product or work for the manufacturer. I got three different answers with little in the way of explanation.
The problem is ongoing, we are trying to split the PDC from the file/print server to address the problem. That hasn't been accomplished yet.
Actually, since my software is a perfect duplicate of their software, made by me on my machine, how can they own it? I owned the blank CD, and all the equipment reading the "CD they own".
I expect that any day now she might start talking to me again.
Well we know those calculations are off.
Everyone's talking about Certifications and Degrees when what he wants is a job. I recommend really polishing up your resume, get several people to look at it and pick it apart for you. It should be no more then 2 pages and preferably one page. I like to stick in a skills section for jobs that might be data mining for key words.
Step two is to get your resume to as many head hunters as possible, this can bite you in the ass if you are applying to a job that has already received your resume from a contract company, but in this entry level position it is an acceptable risk. Head hunters are good because once you add their embellishment to your embellishment you look damn good. They also have short term jobs that allow you to gain experience quickly.
Here is the most important part. Customize your cover letter for every job you apply to, contact each company you apply to no more then one week after they should have received your resume. Try to speak to a hiring manager and verify they have received your resume and find out when they are interviewing. Ask for an interview, this gives you a chance to clear up any concerns they have with your resume, maybe you asked for too much and can afford to take less, maybe they wonder why you are so cheap (agree to take more).
After your interview keep in touch, send a thank you letter. If you are skilled but inexperienced you are really asking them to take a chance on you. Convince them you are worth it.
I beg to differ with you. I work for a Mental Health company with about 250 employees. We have one site with about 100 users, 5 sites w/ 20+/- users, and over a dozen sites with less then 3 users. We have a three person IT department.
There is no way this would work without thin client. Most of our locations use either a direct T1 or residential internet service w/ an IPsec VPN. Some of our sites need computers to handle a dynamic VPN. These sites are way more trouble then the sites with thin client terminals. Our machines are configured to be easily switched out with no data stored locally, but administering windows locally is still a pain.
I have tried linux as the thin client OS, but VPNC only runs under root, I've managed to get it to connect w/ scripts and sudo, but it's unstable.
There are actually ways to take over people's phones and place long distance calls on their dime. Usually these are associated with PBX type systems.
That's actually an example of the doctor trying to screw you and the insurance company stepping in and saying, "Step-off, this is my bitch!". It pisses me off to see the statements I get where the insurance company indicates what they were charged and what they paid. My wife just had some tests that were I think around $112 & $70. Insurance paid around $12 & $17.
That's bullshit!
I haven't seen it here yet, but I use NearlyFreeSpeech.net. They now offer their own domain registration service. From their ideals listed on the page. I can't imagine them caving to any bogus threats. Their prices and support are also great!
satisfied customer
I use JIRA. It's nice and very customizable. I had to change some things since it is very software centric and my issues are network and hardware related. It is good for tracking entire projects and individual trouble tickets.
If I could add one thing it would be the ability to auto-open and close a simple ticket for things like password resets and printer installs.
I beg to differ. I'm in a department of three, director, admin (me), & tech. We support about 200 users at 17 different sites. Only one site has more then 30 users. and more then half only have 2 users. If we didn't use thin client we would need at least double, if not triple the staff. We do use some full PC's, mostly laptops. They are used like a terminal to access citrix, nothing is stored locally.
Some things need work, but all in all it's a remarkably cost effective solution.
I had a similar situation with a prospective employer several years ago. It was a contract company (starts with a k and rhymes w/ dforce). The recruiter offered me a benefits package which turned out to be their internal employee benefit package, not their contract worker benefit package. We went back and forth, they offered to give me the better benefits and cut my hourly rate (no way!).
Finally they agreed to honor the original offer and I had them send me the paperwork. I sat on the paperwork until I heard from another company I was waiting on. The gave me a better offer and I called the orginal recruiter and politily declined a couple days before I was scheduled to start.
I'm probably blacklisted, but realistically I was afraid they would replace me immedietly since I was obviously costing them more then the usual going rate.