If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
|
Thread Tools | Display Modes |
#1
|
|||
|
|||
Bloody heat ?
Nutcracker,
You are the expert, just how hot can the ambient temperature be with theses servers and raid enclosures be before something starts crapping out? We lost the HVAC to the server room, we have doors open (NO windows in room) multiple fans going and it's still hot as hell. Hopefully tomorrow will have it fixed, but this heat concerns me. I had backups to what I thought was everything BUT the cooling. For some reason I did not take that failure into consideration. Thanks, Phil |
#2
|
|||
|
|||
Hello Phil,
To an extent, it depends on the models of the servers. I have never had a problem with storage shutting down, and I have had that stuff running in some very HOT environments. There are parts to my answer. First, most of the systems will go into Thermal Shutdown around 105 degree's F (internal temp). They will shut down (gracefully) until they cool off. At 5 minute intervals, the servers power back on to see if they have cooled down enough to power up and start running again. As long as you have *some* airflow going into the front of those machines, and there is ample airflow to exhaust the heat that comes out of the backs, you should be OK until you get the HVAC system repaired. Second, setting up the servers for a Thermal ASR (automatic server reset) is an OPTION. It doesnt have to be enabled for the servers to operate. Of course, you do so at your own risk. If you have the system management hompage installed on your amchines, then you should be able go into it and tell the server whether or not you want Thermal Reset (or something like that) enabled. If you have any other quetsions, I will be happy to answer them for you. - LC "Phil" wrote in message ... Nutcracker, You are the expert, just how hot can the ambient temperature be with theses servers and raid enclosures be before something starts crapping out? We lost the HVAC to the server room, we have doors open (NO windows in room) multiple fans going and it's still hot as hell. Hopefully tomorrow will have it fixed, but this heat concerns me. I had backups to what I thought was everything BUT the cooling. For some reason I did not take that failure into consideration. Thanks, Phil |
#3
|
|||
|
|||
LC,
That 105 may be the magic number, but that room was allot hotter than that. I'm using about 25 servers, 1850R's and dl380's with a few more odd balls. 10 of the EU's 12 drives ea and 4 newer arrays 14 dives ea. Add the switches, kvm, kvm extenders, ip control, multiple 3, and 6, KVA UPS's, satellite equipment, routers, phone equipment, alarm equipment, CCTV equipment, monitors and lighting and that's allot of heat. There is also a bank of transformers in a big steel cabinet that generate heat too, can hardly touch them, they tie into the panels and backup generators. I didn't realize it until I started counting, but over 200 scsi drives in that mess. Nowhere else do I have a concentration of over about 5 computers. Already switched all my monitors out to flat screens which saved a few bucks on power and heat. We cut a 5-foot hole in the roof of that small building, pulled the drop ceiling tiles out above the racks and temporarily set a 5-foot ceiling fan over the cutout in the roof. That has dropped the temp considerably, and is moving a hurricane like flow of air through now. So far nothing has shut down from heat (that I know of), but this morning I had 6 bix box fans blowing directly on the racks. I had already shut some stuff down, but some stuff has to operate regardless. Heat, and power is one reason I was thinking about upgrading to more modern servers, but I question what dependability I would gain with the new cheaper built equipment. This stuff has worked trouble free, and does everything I need, just not cheap to operate and probably more dependable than new equipment. I thought I had it set up so that any given failure would not stop any process but 95 degrees outside is not helping. Damn, it's been hot. Time to redesign my layout after this. I've learned allot about server operation, but it has been a slow evolution and a continuous learning experience, most the hard way. We never planed to have that much "stuff" in there, it's been a slow process of adding equipment, so it simply wasn't designed for. Thanks again, for the time, Phil "Nut Cracker" wrote in message ... Hello Phil, To an extent, it depends on the models of the servers. I have never had a problem with storage shutting down, and I have had that stuff running in some very HOT environments. There are parts to my answer. First, most of the systems will go into Thermal Shutdown around 105 degree's F (internal temp). They will shut down (gracefully) until they cool off. At 5 minute intervals, the servers power back on to see if they have cooled down enough to power up and start running again. As long as you have *some* airflow going into the front of those machines, and there is ample airflow to exhaust the heat that comes out of the backs, you should be OK until you get the HVAC system repaired. Second, setting up the servers for a Thermal ASR (automatic server reset) is an OPTION. It doesnt have to be enabled for the servers to operate. Of course, you do so at your own risk. If you have the system management hompage installed on your amchines, then you should be able go into it and tell the server whether or not you want Thermal Reset (or something like that) enabled. If you have any other quetsions, I will be happy to answer them for you. - LC "Phil" wrote in message ... Nutcracker, You are the expert, just how hot can the ambient temperature be with theses servers and raid enclosures be before something starts crapping out? We lost the HVAC to the server room, we have doors open (NO windows in room) multiple fans going and it's still hot as hell. Hopefully tomorrow will have it fixed, but this heat concerns me. I had backups to what I thought was everything BUT the cooling. For some reason I did not take that failure into consideration. Thanks, Phil |
#4
|
|||
|
|||
LS, yes I know what I have, UE, arrays !, my daughter is in Europe and
guess that was somewhat on my mind ( EU) Long hot aggravating day. A total loss of time. Phil |
#5
|
|||
|
|||
hehe ... nice.
"Phil" wrote in message ... LS, yes I know what I have, UE, arrays !, my daughter is in Europe and guess that was somewhat on my mind ( EU) Long hot aggravating day. A total loss of time. Phil |
#6
|
|||
|
|||
Interesting you should bring this up, I am currently working for a
company that is suffering through to an A/C failure for the entire building that has the server room shut down. Some joker brought in an old gas station thermometer that was scaled to 135(F) degrees. It topped out in under 15 minutes. As a temp solution we have moved the servers out and distributed them around the network in what has to be one of the most physically unsecured environments possible. Two other sites I've worked at have had heat issues and in both cases the key was air flow just as you have found out. In both instances heat was not considered during the initial establishment of the server room. Here are the two long term solutions they came up with and the solution I have put forward for my current position. 1. Location, one story office complex with server room in a windowless former storage room. Just like you we cut a large hole out, but in the back wall. We then had installed an independent A/C unit (the same as some cheap motels use) with a thermal switch set at 80 degrees. Setting was found through trial and error as we noted the A/C unit needed a bit of a head start to keep the server room temperature below 90 degrees. Realizing that the ambient air temperature needed to be cooler than the max operating temperature of the servers to provide the necessary heat transfer as it was pulled through the cases. 2. Location, at the top of a stair well in a split level building. Server room has it's own HVAC unit, but no back up. The HVAC was running 24/7/365 and the room temperature averaged around 80 degrees regardless of the season or the thermostats settings. Due to power availability and structural issues, we cut a hole in the ceiling and installed a large drum fan with ducting to carry the hot air out of the building and provide a weather proof system. Site is in a hurricane prone area and to date the ducting has kept the rain out. The only draw back is that the fan is very noisy, but then it is an emergency fall back system so it is not on that often. Later we redirected an HVAC vent from a storage closet on the first floor into the server room as a back up. The fan kept the temperature within the operational range, but the front office didn't like the techs working in shorts and t-shirts. 3. Location, former common/meeting room in the center (20 feet from nearest accessible outside wall) of a multi-story office building. They now have some high efficiency A/C unit for home/office room additions that only require a 3 inch opening to connect the inside and outside units. Am proposing we install same. Local climate control folks say that the piping can be run with a bit of custom work on their part. With proper insulation there should be no problem with condensation and the unit can support the expected heat load. This will make the server room independent of the buildings HVAC should it fail in the future. One advantage the climate control folks mentioned is that the replacement unit for the HVAC unit would need to be scaled up as the original unit was never intended to support the sort of heat load that the server room adds to the building. Cost wise it comes out about even and should the server room system fail in the future we have the option to open up both ends of the server room and sponge off the buildings ambient air until it can be fixed. As a point of minor interest, the local community college uses it's former mainframe server room as a mortuary. They long ago built a "Technology Center" to hold their server rooms and computer labs. Until they began teaching "Mortuary Science" the room was of little use due to the custom HVAC system being to cold for classrooms or offices and the being no heating system, the heat from the servers was ample enough. KC "Phil" wrote in message ... LC, That 105 may be the magic number, but that room was allot hotter than that. I'm using about 25 servers, 1850R's and dl380's with a few more odd balls. 10 of the EU's 12 drives ea and 4 newer arrays 14 dives ea. Add the switches, kvm, kvm extenders, ip control, multiple 3, and 6, KVA UPS's, satellite equipment, routers, phone equipment, alarm equipment, CCTV equipment, monitors and lighting and that's allot of heat. There is also a bank of transformers in a big steel cabinet that generate heat too, can hardly touch them, they tie into the panels and backup generators. I didn't realize it until I started counting, but over 200 scsi drives in that mess. Nowhere else do I have a concentration of over about 5 computers. Already switched all my monitors out to flat screens which saved a few bucks on power and heat. We cut a 5-foot hole in the roof of that small building, pulled the drop ceiling tiles out above the racks and temporarily set a 5-foot ceiling fan over the cutout in the roof. That has dropped the temp considerably, and is moving a hurricane like flow of air through now. So far nothing has shut down from heat (that I know of), but this morning I had 6 bix box fans blowing directly on the racks. I had already shut some stuff down, but some stuff has to operate regardless. Heat, and power is one reason I was thinking about upgrading to more modern servers, but I question what dependability I would gain with the new cheaper built equipment. This stuff has worked trouble free, and does everything I need, just not cheap to operate and probably more dependable than new equipment. I thought I had it set up so that any given failure would not stop any process but 95 degrees outside is not helping. Damn, it's been hot. Time to redesign my layout after this. I've learned allot about server operation, but it has been a slow evolution and a continuous learning experience, most the hard way. We never planed to have that much "stuff" in there, it's been a slow process of adding equipment, so it simply wasn't designed for. Thanks again, for the time, Phil "Nut Cracker" wrote in message ... Hello Phil, To an extent, it depends on the models of the servers. I have never had a problem with storage shutting down, and I have had that stuff running in some very HOT environments. There are parts to my answer. First, most of the systems will go into Thermal Shutdown around 105 degree's F (internal temp). They will shut down (gracefully) until they cool off. At 5 minute intervals, the servers power back on to see if they have cooled down enough to power up and start running again. As long as you have *some* airflow going into the front of those machines, and there is ample airflow to exhaust the heat that comes out of the backs, you should be OK until you get the HVAC system repaired. Second, setting up the servers for a Thermal ASR (automatic server reset) is an OPTION. It doesnt have to be enabled for the servers to operate. Of course, you do so at your own risk. If you have the system management hompage installed on your amchines, then you should be able go into it and tell the server whether or not you want Thermal Reset (or something like that) enabled. If you have any other quetsions, I will be happy to answer them for you. - LC "Phil" wrote in message ... Nutcracker, You are the expert, just how hot can the ambient temperature be with theses servers and raid enclosures be before something starts crapping out? We lost the HVAC to the server room, we have doors open (NO windows in room) multiple fans going and it's still hot as hell. Hopefully tomorrow will have it fixed, but this heat concerns me. I had backups to what I thought was everything BUT the cooling. For some reason I did not take that failure into consideration. Thanks, Phil |
#7
|
|||
|
|||
LC, Kevin,
In my case equipment is in a small (1500sf) separate concrete block building with foam insulated, and sheet rocked walls and no windows. We have a 4-foot space between the drop ceiling and the roof. Roof is flat so I thought cutting rafters would be easier, cheaper, and quicker than walls. I'm going to leave that fan on the roof, have a weatherproof enclosure built over the fan with louvers and install steel bars for security. It's a temporary solution, but in light of a total loss of operation, it is working. I'm up in the mountains of North Carolina and we rarely see these high outside temperatures and can hardly justify installing two HVAC systems. (perhaps I need to reevaluate that too) Our compressor took a nosedive and it wasn't a simple quick fix. I've found for the most part, build a backup to the primary "whatever" and the primary won't fail or don't prepare and look out. I'll leave that fan on the roof, just in case, but it is working (cross fingers) and nothing is shutting down now, although still too warm to comfortably work in there with clothes on! Yesterday, prior to the hole/fan we did start having failures in some servers. Luckily it was not all servers and the amazing thing was the their respective backups kept going so I didn't loose one file. I have too much equipment to temporarily relocate it. Our operation is not just file servers We pull data from various satellites process it through multiple machines, while archiving some, it's also routed to multiple other machines. Data routing is very intertwined among various processes as to make a "system", not just a collection of file servers that I could distribute on the network. I did like most of do in these situations, improvised. I will say LC while some servers may shut down at 105, with a high, hurricane like volume of air moving through they seem to hang in there. Just fans on the racks and fans in the buildings open doors did not help much, just more heat than we could push out the doors. The only reason I outline my "mountain" solution, it keeps us going until the primary problem is resolved. We had that ceiling cut open and a big fan setting on the hole in under 3 hours. I'm sure it will take days to weatherproof and secure it. Wow, who would ever thought all that is involved to just keep a computer running! I guess the lesson here is Plan ahead, or you might get caught with your pants down, and loose your shirt, literally, and not with the new secretary ! Stay cool, Phil Oh, By the way LC, that fellow contacted me directly about fedora on the older compact servers. I gave him the info to get him fixed up and he has it going with no problems in graphics mode. If anyone needs to load fedora on any of the older Compaq servers I'll be happy to share the tricks. Have a good day, Phil "Kevin Childers" wrote in message ... Interesting you should bring this up, I am currently working for a company that is suffering through to an A/C failure for the entire building that has the server room shut down. Some joker brought in an old gas station thermometer that was scaled to 135(F) degrees. It topped out in under 15 minutes. As a temp solution we have moved the servers out and distributed them around the network in what has to be one of the most physically unsecured environments possible. Two other sites I've worked at have had heat issues and in both cases the key was air flow just as you have found out. In both instances heat was not considered during the initial establishment of the server room. Here are the two long term solutions they came up with and the solution I have put forward for my current position. 1. Location, one story office complex with server room in a windowless former storage room. Just like you we cut a large hole out, but in the back wall. We then had installed an independent A/C unit (the same as some cheap motels use) with a thermal switch set at 80 degrees. Setting was found through trial and error as we noted the A/C unit needed a bit of a head start to keep the server room temperature below 90 degrees. Realizing that the ambient air temperature needed to be cooler than the max operating temperature of the servers to provide the necessary heat transfer as it was pulled through the cases. 2. Location, at the top of a stair well in a split level building. Server room has it's own HVAC unit, but no back up. The HVAC was running 24/7/365 and the room temperature averaged around 80 degrees regardless of the season or the thermostats settings. Due to power availability and structural issues, we cut a hole in the ceiling and installed a large drum fan with ducting to carry the hot air out of the building and provide a weather proof system. Site is in a hurricane prone area and to date the ducting has kept the rain out. The only draw back is that the fan is very noisy, but then it is an emergency fall back system so it is not on that often. Later we redirected an HVAC vent from a storage closet on the first floor into the server room as a back up. The fan kept the temperature within the operational range, but the front office didn't like the techs working in shorts and t-shirts. 3. Location, former common/meeting room in the center (20 feet from nearest accessible outside wall) of a multi-story office building. They now have some high efficiency A/C unit for home/office room additions that only require a 3 inch opening to connect the inside and outside units. Am proposing we install same. Local climate control folks say that the piping can be run with a bit of custom work on their part. With proper insulation there should be no problem with condensation and the unit can support the expected heat load. This will make the server room independent of the buildings HVAC should it fail in the future. One advantage the climate control folks mentioned is that the replacement unit for the HVAC unit would need to be scaled up as the original unit was never intended to support the sort of heat load that the server room adds to the building. Cost wise it comes out about even and should the server room system fail in the future we have the option to open up both ends of the server room and sponge off the buildings ambient air until it can be fixed. As a point of minor interest, the local community college uses it's former mainframe server room as a mortuary. They long ago built a "Technology Center" to hold their server rooms and computer labs. Until they began teaching "Mortuary Science" the room was of little use due to the custom HVAC system being to cold for classrooms or offices and the being no heating system, the heat from the servers was ample enough. KC "Phil" wrote in message ... LC, That 105 may be the magic number, but that room was allot hotter than that. I'm using about 25 servers, 1850R's and dl380's with a few more odd balls. 10 of the EU's 12 drives ea and 4 newer arrays 14 dives ea. Add the switches, kvm, kvm extenders, ip control, multiple 3, and 6, KVA UPS's, satellite equipment, routers, phone equipment, alarm equipment, CCTV equipment, monitors and lighting and that's allot of heat. There is also a bank of transformers in a big steel cabinet that generate heat too, can hardly touch them, they tie into the panels and backup generators. I didn't realize it until I started counting, but over 200 scsi drives in that mess. Nowhere else do I have a concentration of over about 5 computers. Already switched all my monitors out to flat screens which saved a few bucks on power and heat. We cut a 5-foot hole in the roof of that small building, pulled the drop ceiling tiles out above the racks and temporarily set a 5-foot ceiling fan over the cutout in the roof. That has dropped the temp considerably, and is moving a hurricane like flow of air through now. So far nothing has shut down from heat (that I know of), but this morning I had 6 bix box fans blowing directly on the racks. I had already shut some stuff down, but some stuff has to operate regardless. Heat, and power is one reason I was thinking about upgrading to more modern servers, but I question what dependability I would gain with the new cheaper built equipment. This stuff has worked trouble free, and does everything I need, just not cheap to operate and probably more dependable than new equipment. I thought I had it set up so that any given failure would not stop any process but 95 degrees outside is not helping. Damn, it's been hot. Time to redesign my layout after this. I've learned allot about server operation, but it has been a slow evolution and a continuous learning experience, most the hard way. We never planed to have that much "stuff" in there, it's been a slow process of adding equipment, so it simply wasn't designed for. Thanks again, for the time, Phil "Nut Cracker" wrote in message ... Hello Phil, To an extent, it depends on the models of the servers. I have never had a problem with storage shutting down, and I have had that stuff running in some very HOT environments. There are parts to my answer. First, most of the systems will go into Thermal Shutdown around 105 degree's F (internal temp). They will shut down (gracefully) until they cool off. At 5 minute intervals, the servers power back on to see if they have cooled down enough to power up and start running again. As long as you have *some* airflow going into the front of those machines, and there is ample airflow to exhaust the heat that comes out of the backs, you should be OK until you get the HVAC system repaired. Second, setting up the servers for a Thermal ASR (automatic server reset) is an OPTION. It doesnt have to be enabled for the servers to operate. Of course, you do so at your own risk. If you have the system management hompage installed on your amchines, then you should be able go into it and tell the server whether or not you want Thermal Reset (or something like that) enabled. If you have any other quetsions, I will be happy to answer them for you. - LC "Phil" wrote in message ... Nutcracker, You are the expert, just how hot can the ambient temperature be with theses servers and raid enclosures be before something starts crapping out? We lost the HVAC to the server room, we have doors open (NO windows in room) multiple fans going and it's still hot as hell. Hopefully tomorrow will have it fixed, but this heat concerns me. I had backups to what I thought was everything BUT the cooling. For some reason I did not take that failure into consideration. Thanks, Phil |
#8
|
|||
|
|||
As long as it works for you and the price is right, it's a good
solution. Air flow around the servers is the key. Each server can cool it's self as long as the server room's ambient air temperature is kept cool enough for the server dump it's heat load into it without creating a rising temperature spiral. I was in Fayetteville NC for the first two situations I mentioned. Due to geography, the weather there is not so kind in the summer. To low to get any benefit from elevational cooling but low enough to get fog and high humidity from the mixed airflow off the mountains and the ocean. Yet, to far from the coast to get any true onshore effect. Fortunately the sand hills drain well and are high enough to minimized storm flooding. Best of luck KC "Phil" wrote in message ... LC, Kevin, In my case equipment is in a small (1500sf) separate concrete block building with foam insulated, and sheet rocked walls and no windows. We have a 4-foot space between the drop ceiling and the roof. Roof is flat so I thought cutting rafters would be easier, cheaper, and quicker than walls. I'm going to leave that fan on the roof, have a weatherproof enclosure built over the fan with louvers and install steel bars for security. It's a temporary solution, but in light of a total loss of operation, it is working. I'm up in the mountains of North Carolina and we rarely see these high outside temperatures and can hardly justify installing two HVAC systems. (perhaps I need to reevaluate that too) Our compressor took a nosedive and it wasn't a simple quick fix. I've found for the most part, build a backup to the primary "whatever" and the primary won't fail or don't prepare and look out. I'll leave that fan on the roof, just in case, but it is working (cross fingers) and nothing is shutting down now, although still too warm to comfortably work in there with clothes on! Yesterday, prior to the hole/fan we did start having failures in some servers. Luckily it was not all servers and the amazing thing was the their respective backups kept going so I didn't loose one file. I have too much equipment to temporarily relocate it. Our operation is not just file servers We pull data from various satellites process it through multiple machines, while archiving some, it's also routed to multiple other machines. Data routing is very intertwined among various processes as to make a "system", not just a collection of file servers that I could distribute on the network. I did like most of do in these situations, improvised. I will say LC while some servers may shut down at 105, with a high, hurricane like volume of air moving through they seem to hang in there. Just fans on the racks and fans in the buildings open doors did not help much, just more heat than we could push out the doors. The only reason I outline my "mountain" solution, it keeps us going until the primary problem is resolved. We had that ceiling cut open and a big fan setting on the hole in under 3 hours. I'm sure it will take days to weatherproof and secure it. Wow, who would ever thought all that is involved to just keep a computer running! I guess the lesson here is Plan ahead, or you might get caught with your pants down, and loose your shirt, literally, and not with the new secretary ! Stay cool, Phil Oh, By the way LC, that fellow contacted me directly about fedora on the older compact servers. I gave him the info to get him fixed up and he has it going with no problems in graphics mode. If anyone needs to load fedora on any of the older Compaq servers I'll be happy to share the tricks. Have a good day, Phil "Kevin Childers" wrote in message ... Interesting you should bring this up, I am currently working for a company that is suffering through to an A/C failure for the entire building that has the server room shut down. Some joker brought in an old gas station thermometer that was scaled to 135(F) degrees. It topped out in under 15 minutes. As a temp solution we have moved the servers out and distributed them around the network in what has to be one of the most physically unsecured environments possible. Two other sites I've worked at have had heat issues and in both cases the key was air flow just as you have found out. In both instances heat was not considered during the initial establishment of the server room. Here are the two long term solutions they came up with and the solution I have put forward for my current position. 1. Location, one story office complex with server room in a windowless former storage room. Just like you we cut a large hole out, but in the back wall. We then had installed an independent A/C unit (the same as some cheap motels use) with a thermal switch set at 80 degrees. Setting was found through trial and error as we noted the A/C unit needed a bit of a head start to keep the server room temperature below 90 degrees. Realizing that the ambient air temperature needed to be cooler than the max operating temperature of the servers to provide the necessary heat transfer as it was pulled through the cases. 2. Location, at the top of a stair well in a split level building. Server room has it's own HVAC unit, but no back up. The HVAC was running 24/7/365 and the room temperature averaged around 80 degrees regardless of the season or the thermostats settings. Due to power availability and structural issues, we cut a hole in the ceiling and installed a large drum fan with ducting to carry the hot air out of the building and provide a weather proof system. Site is in a hurricane prone area and to date the ducting has kept the rain out. The only draw back is that the fan is very noisy, but then it is an emergency fall back system so it is not on that often. Later we redirected an HVAC vent from a storage closet on the first floor into the server room as a back up. The fan kept the temperature within the operational range, but the front office didn't like the techs working in shorts and t-shirts. 3. Location, former common/meeting room in the center (20 feet from nearest accessible outside wall) of a multi-story office building. They now have some high efficiency A/C unit for home/office room additions that only require a 3 inch opening to connect the inside and outside units. Am proposing we install same. Local climate control folks say that the piping can be run with a bit of custom work on their part. With proper insulation there should be no problem with condensation and the unit can support the expected heat load. This will make the server room independent of the buildings HVAC should it fail in the future. One advantage the climate control folks mentioned is that the replacement unit for the HVAC unit would need to be scaled up as the original unit was never intended to support the sort of heat load that the server room adds to the building. Cost wise it comes out about even and should the server room system fail in the future we have the option to open up both ends of the server room and sponge off the buildings ambient air until it can be fixed. As a point of minor interest, the local community college uses it's former mainframe server room as a mortuary. They long ago built a "Technology Center" to hold their server rooms and computer labs. Until they began teaching "Mortuary Science" the room was of little use due to the custom HVAC system being to cold for classrooms or offices and the being no heating system, the heat from the servers was ample enough. KC "Phil" wrote in message ... LC, That 105 may be the magic number, but that room was allot hotter than that. I'm using about 25 servers, 1850R's and dl380's with a few more odd balls. 10 of the EU's 12 drives ea and 4 newer arrays 14 dives ea. Add the switches, kvm, kvm extenders, ip control, multiple 3, and 6, KVA UPS's, satellite equipment, routers, phone equipment, alarm equipment, CCTV equipment, monitors and lighting and that's allot of heat. There is also a bank of transformers in a big steel cabinet that generate heat too, can hardly touch them, they tie into the panels and backup generators. I didn't realize it until I started counting, but over 200 scsi drives in that mess. Nowhere else do I have a concentration of over about 5 computers. Already switched all my monitors out to flat screens which saved a few bucks on power and heat. We cut a 5-foot hole in the roof of that small building, pulled the drop ceiling tiles out above the racks and temporarily set a 5-foot ceiling fan over the cutout in the roof. That has dropped the temp considerably, and is moving a hurricane like flow of air through now. So far nothing has shut down from heat (that I know of), but this morning I had 6 bix box fans blowing directly on the racks. I had already shut some stuff down, but some stuff has to operate regardless. Heat, and power is one reason I was thinking about upgrading to more modern servers, but I question what dependability I would gain with the new cheaper built equipment. This stuff has worked trouble free, and does everything I need, just not cheap to operate and probably more dependable than new equipment. I thought I had it set up so that any given failure would not stop any process but 95 degrees outside is not helping. Damn, it's been hot. Time to redesign my layout after this. I've learned allot about server operation, but it has been a slow evolution and a continuous learning experience, most the hard way. We never planed to have that much "stuff" in there, it's been a slow process of adding equipment, so it simply wasn't designed for. Thanks again, for the time, Phil "Nut Cracker" wrote in message ... Hello Phil, To an extent, it depends on the models of the servers. I have never had a problem with storage shutting down, and I have had that stuff running in some very HOT environments. There are parts to my answer. First, most of the systems will go into Thermal Shutdown around 105 degree's F (internal temp). They will shut down (gracefully) until they cool off. At 5 minute intervals, the servers power back on to see if they have cooled down enough to power up and start running again. As long as you have *some* airflow going into the front of those machines, and there is ample airflow to exhaust the heat that comes out of the backs, you should be OK until you get the HVAC system repaired. Second, setting up the servers for a Thermal ASR (automatic server reset) is an OPTION. It doesnt have to be enabled for the servers to operate. Of course, you do so at your own risk. If you have the system management hompage installed on your amchines, then you should be able go into it and tell the server whether or not you want Thermal Reset (or something like that) enabled. If you have any other quetsions, I will be happy to answer them for you. - LC "Phil" wrote in message ... Nutcracker, You are the expert, just how hot can the ambient temperature be with theses servers and raid enclosures be before something starts crapping out? We lost the HVAC to the server room, we have doors open (NO windows in room) multiple fans going and it's still hot as hell. Hopefully tomorrow will have it fixed, but this heat concerns me. I had backups to what I thought was everything BUT the cooling. For some reason I did not take that failure into consideration. Thanks, Phil |
#9
|
|||
|
|||
On Mon, 25 Jul 2005 16:59:43 -0400, Phil wrote:
Nutcracker, You are the expert, just how hot can the ambient temperature be with theses servers and raid enclosures be before something starts crapping out? We lost the HVAC to the server room, we have doors open (NO windows in room) multiple fans going and it's still hot as hell. Hopefully tomorrow will have it fixed, but this heat concerns me. I had backups to what I thought was everything BUT the cooling. For some reason I did not take that failure into consideration. Thanks, Phil Although your servers may not fail right now, running them at a higher tempature stresses them quite a lot. You may see higher than usual drive failures, etc, for the next few months. If you were thinking of buying new equipment, you should probably do so sooner rather than later. Aidan Grey |
#10
|
|||
|
|||
Keep in mind that the temerature of you server room has to be lower at higher altitudes to get the same cooling. Altitude derating factor: FEET METERS COOLING: 0 0 100% 1,000 3,000 95% 1,500 5,000 90% 2,000 7,000 86% 3,000 10,000 80% 3,500 12,000 75% |
|
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Which is better, heat pipe or heat sinks/fan CPU cooling solution. ? | milleron | Asus Motherboards | 1 | May 27th 05 06:37 PM |
Which is better, heat pipe or heat sinks/fan CPU cooling solution. ? | stubborn | Homebuilt PC's | 0 | May 27th 05 12:24 AM |
R9800NP - overclocking problem | power_ranger | Overclocking | 20 | April 19th 04 08:08 AM |
Opteron Overclocking? | Adrian Richards | Overclocking AMD Processors | 9 | October 5th 03 03:20 PM |
XPS, the heat and fan noise problem | Don | Dell Computers | 8 | July 12th 03 08:57 AM |